Skip to main content

Full text of "Tropical American Myrtaceae; notes on generic concepts and descriptions of previously unrecognized species"

See other formats












Published by 

NOVEMBER 30, 1956 





Professor of Botany, University of Michigan 



Published by 

NOVEMBER 30, 1956 



Tropical American Myrtaceae 

The following notes have been prepared as a general introduction 
to a formal treatment of the Myrtaceae of Peru which is now in the 
course of preparation. A large part of the present paper is devoted 
to the characterization and description of approximately 80 species 
and sub-specific taxa, mostly from Peru, which appear to be new to 
science. The keys which are set forth below are intended primarily 
for the student who wishes to place the newly described species 
properly among their congeners or to follow the lines of reasoning 
by which I became convinced that the specimens represented 
undescribed taxa. Both keys and descriptions have been abridged 
from those which were originally prepared for the Flora of Peru, so 
that extra-Peruvian species have not been included in the keys 
unless there seems reason to believe that ultimately they may be 
found in that country. 

The American representatives of the Myrtaceae have long been 
considered a "difficult" group, and one in need of much systematic 
study. This is partly because of the undoubtedly very large number 
of species in the American tropics, and partly because of certain 
features inherent in the plants themselves (e.g., flowers and fruits 
of the same species are rarely obtainable at the same time). Dis- 
tinctions between species and even between genera are sometimes 
nice ones, for flowers as well as vegetative structures are relatively 
uniform throughout the whole family. Identification of flowering 
material is difficult for the casual student because distinctions 
between genera have been made primarily on characters of the 
mature embryo. Embryo characters are not always easy to observe 
in dried material, even in specimens with mature fruit, and they are 
impossible to make out from flowering specimens. 

One of the principal impediments in the way of serious work on 
the American Myrtaceae during the last century, paradoxically 
enough, seems to have been the existence of two important and 
indeed monumental treatises by Otto Karl Berg (1815-1866), 



a Professor of Pharmacy at the University of Berlin. 1 Berg laid the 
foundations for all later taxonomic work on this family in the New 
World, but his monographs are difficult to use except after prolonged 
study and collation. Neither one was provided with a workable key 
to species, and the generic keys can be used for specimens in the 
fruiting condition only. Berg's taxonomic philosophy led him to 
believe that most species were sharply restricted in range to single 
phytogeographic regions, which in turn increased inordinately the 
number of species he described as new. 

Probably the greatest obstacle confronting the would-be student 
has been the vast number of species involved. According to Berg 
himself, 2 the known American species in 1859 comprised a total of 
1726, of which he personally had proposed 1008. The largest genera 
were Eugenia (537 species), Aulomyrcia (240), Myrcia (184) and 
Psidium (101). The published enumeration listed 696 species in 
Brazil, with the greatest concentration in the southeastern part of 
that country (390 species in Minas Gerais and Goiaz together, 265 in 
Rio de Janeiro, and 175 in Parana and Sao Paulo). Only 161 species 
were listed from the whole Amazonian region, and only 54 from 
Peru and Bolivia together. Because of the large numbers of species 
in certain genera and in certain areas, and because of the way in 
which Berg organized his monographs, it is for any practical purpose 
impossible to identify an unknown specimen unless one already has 
some knowledge of the Myrtaceae, or unless the specimen is in good 
fruiting condition and comes from an area whence few species are 

Berg's monographs in spite of the difficulties attendant upon 
their use were major contributions to systematic botany. His 
major theoretical work was perhaps the clarification of the taxonomy 
of the tribe Myrteae, which is discussed below. His generic concepts 
were probably somewhat narrower than those of most modern 
workers, but he was a keen and accurate observer. He recognized 
as new a great many species which were indubitably distinct, and 
a great many others which now appear to have been based upon 
rather trivial characters. When one considers that many of the 

1 Revisio Myrtacearum Americae, Linnaea 27: 1-472; index, op. cit. 786-795. 
1855-1856 (also distributed as a separate, without index). Mantissa I. ad re- 
visionem Myrtacearum Americae, op. cit. 29: 207-264. 1858. Mantissa II. ad 
revisionem Myrtacearum Americae, op. cit. 30: 647-713. 1861. Mantissa III. 
ad revisionem Myrtacearum Americae, op. cit. 31: 247-262. 71862. Myrtaceae, 
in Mart. Fl. Bras. 14, pt. 1: 1-656, pis. 1-82. 1857-1859. 

2 Conspectus distributions Myrtacearum Americae hue usque cognitarum, 
Mart. Fl. Bras. 14, pt. 1: 619-622. 1859. 


species recognized by Berg were known to him only through single 
and often imperfect even sterile specimens, and that only rarely 
was he able to see both flowers and fruit in the same species, it is 
at once apparent that the quality of his work was extraordinarily 

When the present work on the Myrtaceae of Peru was begun, it 
was soon apparent that the number of species native within the 
political boundaries of the country would approximate 150. This 
suggested that a large number of the species were new to science, as 
Berg had listed no more than 54 from Peru and Bolivia together, and 
subsequent authors had added hardly more than a score of species 
to the flora of Peru. Another alternative seemed possible, namely, 
that many of the Peruvian species would prove to be identical with 
others which had been described previously, by Berg or by others, 
from extra- Peruvian material. It therefore seemed necessary to 
undertake at least a cursory survey of all tropical American repre- 
sentatives of the family, in order to establish and delimit the princi- 
pal patterns of distribution. 

In the course of this survey I have examined most of the material 
which is available at the United States National Herbarium and at 
Chicago Natural History Museum; the latter is very rich in South 
American Myrtaceae and has many isotypes and type fragments 
from the classical collections. I have also had the privilege of 
studying selected specimens from the New York Botanical Garden; 
the Harvard University herbarium; the Naturhistorisches Museum, 
Vienna; and the herbarium of the Universidad Nacional Mayor de 
San Marcos, Lima. With the aid of a grant from the Horace H. 
Rackham School of Graduate Studies, University of Michigan, I 
was enabled to spend the months of July and August, 1954, in the 
study of these plants as they are represented in European herbaria. 
Most of Berg's types, or duplicates of these, are preserved at Munich 
(in the Brazilian herbarium of Martius), at Geneva (where are also 
the types of the DeCandolle Prodromus), at Paris, or at Brussels. 
Many isotypes are to be found at the British Museum (Natural 
History), or at Kew. I was able to visit each of the above herbaria 
at least for a short time, and I am especially grateful to the authorities 
at Geneva, at Paris, and at the two British herbaria, for permission 
to borrow certain critical material for further study. Although I 
was unable to visit Vienna, Dr. K. H. Rechinger very kindly located 
for me, and forwarded to me for study, most of the unique specimens 
which Berg had described from among the collections there. A few 
of Berg's types which were at Berlin, and apparently not represented 


by duplicates elsewhere, were presumably destroyed in World 
War II. 

The following generalizations relative to range patterns in tropi- 
cal American Myrtaceae are presented as tentative, after study of 
the materials discussed above; it is thought that future work will not 
markedly affect the principal conclusions, but concepts of individual 
species may be radically changed after study of the groups to which 
they belong, and names applied to Peruvian species may on this 
account have to be changed in some instances. 

Distribution of South American Myrtaceae 

1. The Chilean Myrtaceae are all endemic, and none of the 
species ranges as far north as Peru. This is not unexpected, in view 
of the peculiar isolation which Chile enjoys, and it is mentioned here 
chiefly because Berg wrongly attributed a number of Chilean species 
to Peru, and Peruvian species to Chile. This error came about 
because of lack of sufficient knowledge of the itineraries of certain 
collectors, particularly Dombey and Poeppig, who visited both 
countries. The Myrtaceae of Chile are probably better known than 
those of any other area of comparable extent in South America, 
thanks chiefly to the efforts of Eberhard Kausel of Santiago, who 
has published numerous papers on the subject. 1 

2. A majority of the species of southern Brazil (i.e., from south 
of an arbitrary line drawn at about 15 S. Lat., and including all of 
Brazil south of Minas Gerais, southern Goiaz and southern Mato 
Grosso) do not extend northward into the Amazon basin and are 
not likely to be found in Peru. 

3. Species of the Bolivian lowlands, even the lowlands of 
northern Bolivia, have little in common with most Peruvian species 
but show strong affinities with species ranging from southern Brazil 
to Uruguay and northern Argentina. The line between Bolivian 
and Peruvian lowlands appears to be crossed by relatively few 

4. Species from the northeastern states of Brazil (e.g., Bahia, 
Piauhy, and Ceard) are often distinct from those of southern Brazil 

1 Cf. Contribution al estudio de las Mirtaceas Chilenas, Rev. Argent. Agron. 
9: 39-68. 1942; op. cit. 221-243. 1942; op. cit. 11: 320-327. "1944" (1945); Notas 
mirto!6gicas, Lilloa 13: 125-149. "1947" (1948). The first two papers comprise 
a revision of the Chilean genera and species; the third is in the form of a supple- 
ment incorporating some new information. The final paper presents miscel- 
laneous new observations and is concluded with a new key to the Chilean genera. 


and usually seem distinct from those of the Amazon region but may 
show affinities with both. This area is much in need of further 

5. Some species of the Guiana lowlands, and most species of the 
Amazon lowlands, seem potentially wide-ranging and should be 
considered in any treatment of the Peruvian flora. At least a few, 
and probably more, species certainly range widely from the West 
Indies through the Guianas and most of lowland Brazil, Peru 
and Bolivia. 

6. Most of the myrtaceous species found by early collectors 
on the upper Amazon and its tributaries, as far down as Ega (Teffe"), 
have since been re-collected in Peru. A large number of distinctive 
species, however, have been collected along the Rio Negro and other 
rivers which drain into the Amazon from the north but have not 
been collected along the upper reaches of the Amazon itself. It 
appears that many species which range northward into Venezuela 
in the Rio Negro drainage, although they may reach the Amazon 
near the mouth of the Rio Negro, do not extend far up the Amazon 
and are not to be expected in Peru. 

7. Plants of the inter-Andean valleys, and from high elevations 
up to near the limits of vegetation, seem in Peru to be mostly 
endemic, although the Peruvian species are often very similar to 
those of the more northern parts of the Andes. 

In view of the above conclusions I have, in describing the follow- 
ing novelties from Peru and elsewhere, considered in the most careful 
detail all the species previously described from the entire Amazon 
basin ; I have paid close attention to species described from northern 
South America, and more than casual attention to species described 
from eastern and southern Brazil. Further taxonomic studies of 
the very numerous published species from these latter areas, how- 
ever, may bring about changes in concepts and in nomenclature 
which will affect some Peruvian species also. It is further possible 
that a few Peruvian species will prove to be identical with some 
of those described by Linnaeus, Jacquin, Lamarck, and other early 

Generic Concepts 

One of Berg's principal contributions to the taxonomy of the 
Myrtaceae was the documentation of the fact that the tribe Myrteae, 
including practically all the American members of the family, can be 


divided into three coherent groups (subtribes) based on characters 
of the embryo. This had been discerned much earlier by DeCandolle, 
but the specimens available to that author were much too few in 
number to enable him to make the general conclusions which were 
set forth by Berg. Berg recognized three subtribes, which he called 
Myrcioideae, Eugenioideae, and Pimentoideae, and which included, 
respectively, 11, 12, and 17 genera. 1 

More recent workers have not always recognized all the genera 
admitted by Berg. Bentham (Benth. & Hook, f . Gen. PL 1 : 712-720. 
1865) reduced the numbers of genera in the subtribes to 3, 3, and 7, 
respectively. Niedenzu (Nattirl. Pflanzenfam. Ill, pt. 7: 64-86. 
1893) reversed the trend and admitted 5, 4, and 13 genera, re- 
spectively. Conservative modern opinion would agree upon the 
recognition of at least the following generic groups (Chilean genera 
not included) : 

1. MYRCIINAE. Cotyledons foliaceous, contortuplicate; radicle 

(a) Calyptranthes Sw. Buds closed, calyx calyptrate. 

(b) Marlierea Camb. Calyx closed or barely open in bud, split- 
ting irregularly in anthesis. 

(c) Myrcia DC. Calyx with 5 (rarely 4) distinct lobes. 

Other genera proposed by Berg depend upon single characters 
of the anthers (Cerqueiria, Gomidesia) or of the calyx (Aulomyrcia, 
Calycampe, Calyptromyrcia, Eugeniopsis, Rubachia). None of these 
has been much studied or widely accepted except Aulomyrcia, which 
has been taken up in recent years by Miss Amshoff in her studies 
of the Myrtaceae of northern South America. The genus Myrc- 
eugenia Berg, with a disjunct range in Chile and Uruguay, is 
apparently a distinct group. The others mentioned above, however, 
are but doubtfully distinct from Myrcia. Most of the species are 
south-Brazilian, and proper elucidation of the genera must wait 
upon taxonomic studies in this region. 

1 According to the International Code (Article 29), these subtribes are properly 
known as Myrciinae Berg, Eugeniinae Berg, and Pimentinae Berg. The pertinent 
rule states: "When the name of a taxon . . . has been published with an improper 
termination . . . the ending must be changed to accord with the rule, without 
change of authority." The names of these three subtribes were first published 
with the proper termination by Niedenzu (Natiirl. Pflanzenfam. Ill, pt. 7: 62. 
1893) but that author substituted Myrtinae for Pimentinae, contrary to the 
provisions of Article 29. The names Myrciinae and Eugeniinae have usually been 
attributed to Niedenzu, e.g., in Dalla Torre & Harms (Gen. Siphon. 348. 1903), 
but are properly attributed to Berg. 


2. EUGENIINAE. Cotyledons fleshy, distinct or conferruminate 
or the embryo undivided ; radicle very short. 

(a) Eugenia L. Calyx-lobes usually 4, distinct; hypanthium not 
prolonged above the summit of the ovary; ovules numerous. 

(b) Calyeorectes Berg. Buds closed, irregularly splitting; hy- 
panthium prolonged; ovules numerous. 

(c) Myrciaria Berg. Calyx-lobes 4; hypanthium prolonged, cir- 
cumscissile at base; ovules 2 in each locule. 

Certain additional genera, not recognized by Berg, have been 
accepted by Amshoff (e.g., in Fl. Suriname 3: 56-158. 1951). These 

(d) Plinia L. Buds closed or nearly so, the calyx splitting 
irregularly; flowers glomerate, involucrate; hypanthium pro- 
longed; ovules 2 in each locule. 

(e) Catinga Aubl. Calyx-lobes 4, thickened distally and coher- 
ent in the bud; hypanthium not prolonged; ovules numerous. 

Of the remaining genera accepted by Berg, none has been widely 
accepted. Two of these (Phyllocalyx, Stenocalyx) depend upon 
a character of the inflorescence which is easily recognized but of 
doubtful importance. Myrcianthes, according to Bentham (op. cit. 
715), was wrongly placed in the Eugeniinae and belongs rather with 
Myrtus; and Mitranthes, according to Bentham (I.e. 717) belongs 
with Calyptranthes in the Myrciinae. Schizocalyx, Siphoneugena and 
Hexachlamys are based on calyx characters which scarcely serve 
to distinguish the first from Calyeorectes and the last two from 
Eugenia. Acca certainly belongs in the Pimentinae, and is treated 
below. In recent years it has been shown that Aulacocarpus prob- 
ably belongs to the Melastomaceae. 

3. PIMENTINAE. Embryo spiral or curved; radicle elongate; 
cotyledons very short. 

(a) Campomanesia R. & P. Ovary usually 6- to many-locular; 
calyx-lobes 5, membranaceous; embryo spirally involute, the 
testa membranaceous; ovules 2- or rarely 4-seriate. 

(b) Psidium L. Ovary 2- to many-locular; calyx closed in the 
bud or with 5 short lobes; embryo circinate; ovules many- 


(c) Calycolpus Berg. Ovary usually 4- to 5-locular; calyx-lobes 
5, often foliaceous, spreading or reflexed in bud; embryo 

(d) Pimenta Lindl. Ovary bilocular; calyx-lobes 4 or 5; embryo 
cyclic or sub-spiral; ovules 1-6, pendulous; inflorescence 

(e) Ugni Turcz. Calyx-lobes 4 or 5; embryo arcuate; peduncles 
1-flowered; anthers sagittate, the connective dilated. 

Cf) Myrteola Berg. Calyx-lobes usually 4; embryo arcuate; 
peduncles 1-flowered; anthers subrotund, the filaments 

(g) Blepharocalyx Berg. Calyx-lobes 4, ciliate, deciduous; 
embryo sub-spiral or spiral, the testa membranaceous; 
peduncles dichotomous, mostly 3- or more-flowered. 

The rest of the genera accepted by Berg have been little studied. 
The monotypic Paivaea differs from Campomanesia chiefly in the 
patelliform-dilated base of the young calyx. Also monotypic are 
Calyptropsidium and Psidiopsis, which differ from Psidium each 
by a single character of the calyx and which are difficult to defend 
as independent genera. Pseudocaryophyllus was scarcely known 
to Berg except from descriptions and illustrations and was founded 
at least in part upon a misunderstanding of the morphology of one 
species. Amomis has been placed by many authors, including 
Bentham and Niedenzu, in the synonymy of Pimenta. Acrandra 
appears to be distinct by virtue of its apiculate anthers. Britoa, 
Abbevillea and Lacerdaea were placed by Bentham in the synonymy 
of Campomanesia, from which they differ by calyx characters of 
doubtful significance. 

The genus Myrtus is not included in the above list. Berg as- 
signed 34 American species to Myrtus, in addition to 5 species about 
which he had some reservations. At least half the species included 
by Berg in Myrtus have been shown to belong to other genera, and 
the rest have not been critically studied. This matter will be dis- 
cussed below, with reference to the Peruvian genera in particular. 

The Myrtaceous Inflorescence 

Students of the Myrtaceae have to a considerable degree passed 
over the inflorescence in their search for characters which may be 
used in the delimitation of taxa. The subject of inflorescence 


morphology was indeed discussed at some length by Bentham in an 
early paper 1 and before him by Grisebach, but it has been little 
stressed by more recent workers. It is therefore thought to be 
worth while to discuss the myrtaceous inflorescence here, and to 
follow this discussion by a summary of the features which seem to 
be of taxonomic significance. 

Most of the species of the Myrciinae can be recognized as such, 
even in the absence of both flowers and fruit, from the inflorescence 
alone; it is even possible to make many identifications to genera 
on this basis. Most of the American genera of the Eugeniinae and 
Pimentinae can also be recognized by a combination of inflorescence 
characters with calyx characters which are apparent either in 
flowering or in fruiting material. Various markedly different in- 
florescences occur among the American Myrtaceae, but the differ- 
ences are, at least to some extent, superficial. All known types are 
derivable from a single basic pattern. Branching patterns of 
vegetative stems are fundamentally similar to those of flowering and 
fruiting stems, and an understanding of each is essential to an 
understanding of the morphology of the inflorescence. 

In the tribe Myrteae the leaves and branches are normally 
decussate; that is, in pairs at the nodes, with those at any one node 
standing at right angles to those at the nodes above and below it, 
and the leaves at alternate nodes standing directly above or below 
one another. This arrangement prevails not only in simple (un- 
branched) axes, but also in branching systems. Branching at the 
lowest node of an axillary branch, for example, will occur in the 
plane of the tangent at right angles to the plane of branching at the 
node from which the axillary branch arises. The sequence is readily 
observed in leafy branch-systems, in bracteate inflorescences devoid 
of ordinary foliar structures, and in transitions from vegetative to 
reproductive branches. The inflorescence is thus interpreted as an 
axillary branch of which the primary branchlets are normally 
decussate, exactly in the manner of sterile branches. Deviations 
from the usual pattern, e.g., the alternate branching sometimes 
seen in Myrcia near the tips of the panicles, an occasional unilateral 
panicle in Calyptranthes, or occasional solitary flowers in normally 
racemosely flowered species of Eugenia, are assumed to represent 
reduced or derived conditions. 

It is sometimes difficult to recognize the proper limits of a single 
inflorescence. In some genera, e.g., Myrcia, almost any new branch 

1 Notes on Myrtaceae, Journ. Linn. Soc. Bot. 10: 101-166. 1869. 




FIG. 1. Decussate leaves and branches in the Myrtaceae. Large circles 
represent the principal branch; small circles, subsidiary branching axes (i.e., 
a leafy branch, the axis of a raceme, or a pedicel); separate lenticular figures, 
petioles of ordinary leaves; angular lines, bracts. Bracts and bracteoles are in 
solid black. 

arising from the axil of a foliage leaf may represent a potential 
inflorescence. At the one extreme the whole branch may be modified 
in the direction of flower production, all its leaves reduced to minute 
deciduous bracts and its secondary branchlets irregularly disposed; 
such a branch is readily defined as a single inflorescence. The same 
new axillary branch, on the other hand, may bear normal leaves 
at most of the nodes and miniature inflorescences from certain axils 



only (e.g., the lower ones). The partly leafy branch is the homo- 
logue of the completely fertile branch; what, then, represents the 
inflorescence of this species? In practice this can usually be deter- 
mined by correlation with leaves of a certain size, i.e., those occur- 
ring on permanent mature twigs; leaves are often reduced in size 
when occurring on temporary (inflorescence) branches. 

FIG. 2. Schematic representation of flowering in Eugenia. Leafless raceme 
from the lower node; partially fertile branch from the upper node. Bracts and 
bracteoles in solid black. See description in text. 

In other genera, e.g., in Eugenia, the flowers may occur in 
a short leafless raceme (the whole subtended by a foliage leaf), or 
in the same species they may be solitary at the lowest (leafless) 


nodes of a new branch which bears normally developed foliage 
leaves at the upper nodes. In the first case it may be assumed that 
the fertile branch failed to develop beyond the lower nodes; in the 
other case that the axis continued to grow and produce leaves. The 
short leafless raceme is the homologue of the leafy branch with 
flowers at the lowest (and leafless) nodes, but it is difficult to re- 
gard the latter as an "inflorescence." 

Finally, it should be stated that the primitive myrtaceous in- 
florescence is assumed to be one in which the primary axis, and each 
of the subordinate axes, is terminated by a flower. 

The inflorescence (or to put it perhaps more precisely, the 
potentially floriferous branch which develops from the axil of a leaf) 
may be modified in aspect and morphology in one or more of the 
following ways: 

1. Reduction of leaves to small bracts which may be colored or 
scarious, persistent or deciduous. 

2. Abortion of the terminal bud of the primary axis or those 
of the secondary axes. 

3. Reduction in length of the primary axis, with or without 
a corresponding reduction in the number of nodes. 

4. Termination of the primary axis by the formation of a flower 
at the first node (inflorescence a dichasium) or at one of the suc- 
ceeding nodes. 

5. Reduction of secondary axes to 1-flowered bracteolate pe- 
duncles (inflorescence a raceme). 

6. Dehiscence of secondary axes from one or both sides of 
a node, with the production of irregularly branched panicles. 

7. The regular reduction, from the base to the apex of the 
inflorescence, of internode length and number of nodes in all sub- 
ordinate axes. 

In the following paragraphs are described the principal modi- 
fications which result from combinations of the above: 

1. The raceme. This type is characteristic of the subtribe 
Eugeniinae and occurs occasionally in genera of the Pimentinae. 
The raceme is an axillary and usually leafless inflorescence with 
a single primary axis and several or many nodes; the axis is theoreti- 
cally indeterminate but usually aborts at the tip and lacks a terminal 
flower; the internodes may be elongated or much abbreviated; the 



branches from the axis are in decussate pairs, all about the same 
length and reduced to bracteolate "pedicels" terminated each by 
a solitary flower. In exceptional cases the racemes are compound, 
e.g., in Eugenia florida, in which the individual flowers of the 
raceme are wholly or partly replaced by small racemosely flowered 
branches, or in E. stipitata, in which each peduncle may bear a simple 
dichasium instead of a single flower. 

FIG. 3. Modified racemose inflorescences in Eugenia. Left, E. stipitata 
subsp. sororia: the fertile branch is a short raceme in which some of the peduncles 
are 3-flowered. Right, E. punicifolia, in which the flowers are reduced to a single 
pair in each raceme. Bracts and bracteoles in solid black. 

When the axis of the raceme is much reduced in length and the 
nodes are approximate, the number of nodes may be correspondingly 
reduced and the flowers may appear to be umbellate or glomerate. 
The most extreme reduction occurs in species like Eugenia puni- 
cifolia, in which the flowers are a single pair in each axil, one arising 
from each side of the abortive axillary bud. 

In some species of Eugenia, as mentioned above, a new leafy 
branch may have the lowest two or three nodes fertile and bracteate 
rather than leafy, but the upper nodes sterile and normally leafy; 
in such species the flowers may thus appear to be solitary at leafless 
nodes. In the same species, from other leafy nodes, there may arise 
abbreviated racemes in which the leaves are all reduced to small 
bracts and the axis is abortive. 

A similar situation prevails in Campomanesia lineatifolia (Pimen- 
tinae), a species in which the flowers usually occur at the two lowest 


nodes of an otherwise leafy branch but in which the branch may be 
reduced to a short leafless axis comprising only the two fertile nodes. 

In any inflorescence of the racemose type each pedicel is sub- 
tended by a bract. If the number of flowers in a raceme is reduced 
to one, as may rarely happen, it is still ordinarily possible to observe 
the pedicellar bract, or at least the bract-scar, and usually possible 
in addition to observe the abortive terminal bud of the inflorescence. 
In species which normally bear solitary flowers in the axils of 
ordinary leaves (see discussion below), the pedicel bears no bract 
at base and the flower itself represents the termination of the axis. 

2. The myrcioid panicle. With few exceptions this type is 
found in the principal genera of the Myrciinae, with some modi- 
fications as noted below. The axillary panicle is compound, or 
decompound, with the primary axis and each of the subsidiary 
branches terminating in a flower. The principal branches are op- 
posite and decussate, the lower ones elongate and usually themselves 
compound and many-flowered, and the ultimate branchlets of the 
panicle usually with the flowers aggregated in threes (i.e., in simple 
dichasia) near the tips. The transition between elongate basal 
branches and simple dichasia at the tips is accomplished by 
gradual reduction, in successive branches, of the number of nodes 
and the length of the internodes. Branching toward the tips of the 
panicle may be irregular because of abortion of one or both buds 
at a given node, or because of the dehiscence (at the point of origin) 
of one or both branchlets after the initiation of growth; either of 
these processes may result in, or be accompanied by, distortion or 
zigzag growth of the axis which remains. 

Because of the irregular development of the myrcioid panicle 
it is usually not practicable to state the usual number of flowers 
in the panicle of a given species, except within broad limits. A 
large panicle may contain 300 flowers or even more; a small panicle 
in the same species may contain no more than 50. In some species 
the number of flowers may be reduced to 3, 2, or even one. On 
the whole it is a fair generalization that a "many-flowered" species 
will probably prove upon examination to be a member of the 
Myrciinae, whereas among the Eugeniinae an inflorescence with as 
many as 50 flowers is a rarity. 

Usually the myrcioid panicle can be distinguished at a glance 
from a compound racemosely branched panicle (which may be 
"many-flowered," as in Eugenia florida) by the long lower branches. 
In a racemosely branched panicle the pedicels and the internodes 

FIG. 4. Diagram of the myrcioid panicle. Position of flowers and branches 
indicated to the extent necessary to make clear the pattern of branching which 
is repeated in each part of the panicle. Branches in one plane are shown; those 
at right angles to this plane are omitted, but their positions are indicated by 



are essentially of uniform length, so that the flowers appear uni- 
formly distributed in the inflorescence. In the myrcioid panicle, 
especially after anthesis, the observer receives the general impression 
of numerous elongated slender branches which bear short-stalked 
or sessile flowers or flower clusters at irregular intervals, and par- 
ticularly near their tips. 

Among the myrcioid genera in our flora it is possible to recognize 
two tendencies in the branching of the inflorescence. In the genus 
Myrcia, and in the segregate genera Aulomyrcia and Gomidesia, the 
principal axis of the inflorescence is as long as, or a little longer than, 
the primary lateral branches; the panicle, as a result of this, is about 
as long as wide. In Marlierea and Calyptranthes the principal axis 
may abort at the node where the lowest lateral branches emerge, so 
that the inflorescence as a whole appears to consist of paired spikes 
or panicles arising from the same axil. This condition is found 
regularly in some species, occasionally in others, and rarely or 
never in a third group. It is thus by no means consistent in the 
genera in which it occurs, but as far as I am aware it never occurs in 

3. The dichasium. This type characterizes the Pimentinae and 
is found in certain genera of the Eugeniinae (e.g., Anamomis Griseb.). 
The primary axillary axis is immediately determinate; it terminates 
in a flower at the first node. The flowers may be thus normally 
solitary and subtended by normal foliage leaves. Invariably the 
flower is subtended by a pair of opposite bracteoles which indicate 
the potentiality of lateral branching at these points, and in various 
species solitary flowers are often found on the same plant with 
3- or 7-flowered dichasia. 

Branching of the primary axis may take place at the node just 
beneath the terminal flower (i.e., from the axils of the subtending 
bracteoles), with the production of a simple (3-flowered) dichasium. 
By further symmetrical development of the lateral branches, the 
inflorescence may become a 7-, 15- or rarely 31-flowered dichasium. 
In the forks of a compressed 15- or 31-flowered dichasium the ter- 
minal flowers may fail to develop to maturity. 

In most species with predominantly solitary flowers or few- 
flowered dichasia, a majority of the mature leaves bear flowers in 
their axils, and there appears to be no marked restriction of fertility 
to the lower nodes of a shoot. When solitary flowers and dichasia 
occur on the same plant, the most vigorous nodes and leaves appear 


FIG. 5. Details of branching at the tips of the myrcioid panicle. Left, 
a cluster of three nearly sessile buds forming a "false dichasium," in which the 
lateral flowers arise from the second node below the terminal flower; the buds 
at the first node are abortive. Right, another species, in which the lateral flowers 
of the dichasium are borne on slender pedicel-like branches from the node im- 
mediately below the terminal flower. 

FIG. 6. A dichasium in Eugenia. In some species of Eugenia and Psidium, 
solitary flowers and 3-flowered dichasia may be found on the same plant. 

to be correlated with inflorescences bearing the larger numbers of 
flowers. In certain species of Psidium, and perhaps in some species 
of Eugenia also, the dichasia may arise from leafy axils, or in the 
same species from leafless nodes at the bases of leafy shoots; this 
condition is apparently less frequent among species in which the 
flowers are borne in dichasia than among the racemose-flowered 


The Myrtaceous Leaf 

The keys and descriptions which follow include some terms 
which are used repeatedly in certain restricted senses; the following 
notes describe the ways in which these terms apply specifically to 
the Peruvian Myrtaceae: 

The leaves of most American Myrtaceae have a straight midvein 
which extends the entire length of the blade; the leaves are pinnately 
veined, and the midvein is very markedly thicker than any of the 
lateral veins. The midvein almost without exception is conspicuous 
on the lower (dorsal) surface of the leaf, and considerably raised 
above the surface as a convex line; in some species it is elevated to 
such an extent that its diameter at right angles to the leaf is as great 
as the diameter in the plane of the leaf. The midvein on the upper 
(ventral) surface of the leaf may be conspicuous and somewhat 
elevated or at the other extreme nearly invisible. The vein proper 
may be convex, flat, somewhat concave, or with a narrow groove 
or channel running the entire length; in a few cases the midline of 
the vein may be elevated in a narrow ridge. In addition to the above 
modifications in the surface of the vein proper, the whole vein may 
be impressed; that is, depressed in a trough or furrow below the 
general surface of the leaf. In general, therefore, to say that a vein 
is impressed refers to the position of the whole vein in relation to the 
rest of the leaf, not to any modification of the vein itself. 

The lateral veins arise at intervals from the midvein and extend 
toward the margin, forming usually an acute angle with the distal 
portion of the midvein. The lateral veins are said to occur in pairs, 
but this is something of a misnomer, because the veins on opposing 
sides of the midvein are more often alternate than opposite. The 
word "pairs" is used to avoid a longer and more clumsy expression; 
i.e., it is preferable to say "12 pairs" than "12 on each side of the 
midvein." The lateral veins are often nearly unbranched, that is, 
they lack major branches and appear to extend without inter- 
ruption from the midvein to the margin or near it; in some species 
the lateral veins fork in the distal half or the distal third. The 
secondary veins which connect the laterals are often very small and 
appear to the eye as a small network between larger veins. 

As a lateral vein approaches the leaf margin it invariably curves 
toward the apex of the leaf and joins the next succeeding vein. In 
this way are produced various kinds of marginal veins which are 
approximately parallel to the leaf margin. In some species the 


lateral veins diminish gradually in thickness from base to apex and 
recurve markedly toward the midvein before joining the succeeding 
laterals. In this case no marginal vein, as such, can be distinguished, 
and in extreme cases the lateral veins become so much attenuated 
toward the tips that one can with difficulty make out the points of 
union with the succeeding veins. In other species the lateral veins 
are straight and of uniform diameter, but their distal ends are 
connected by a series of arches or loops which together comprise 
a marginal vein; this arched marginal vein may be thinner than the 
laterals, but in many species its diameter is as great as that of the 
laterals. In the species which are thought to present the most 
highly evolved vein pattern, the marginal vein is a characteristic 
and conspicuous feature of the leaf, arching almost not at all between 
the lateral veins and thus nearly straight except that it follows the 
curve of the leaf margin. When the marginal vein is of this kind, 
the laterals may extend directly from midvein to marginal vein and 
enter the latter almost at right angles, with little or no apparent 
tendency to turn toward the apex of the leaf. 

The number of pairs of lateral veins is used in various instances 
to distinguish species. The number is constant within rather broad 
limits only, and the actual number present in a given leaf is often 
doubtful. A precise count can often be made in species which have 
strong rib-like veins and few or no intermediate veins, but such 
species are in the minority. Far more abundant are species in which 
the major pairs of veins are but little stronger than the numerous 
slender intermediates which may curve or branch or join another 
vein. In doubtful cases the difficulty can often be resolved by 
counting the number of arches in the marginal vein; as a general 
rule these arches extend in unbroken curves from one major lateral 
vein to the next, and small intermediate veins cause little or no 
break in the smooth curve of the arch. At best the numbers of veins 
as stated are mere approximations. 

The Genera of Myrtaceae Which Occur in Peru 

The following key is intended as an indication of the taxonomic 
arrangement which is suggested by consideration of the morpho- 
logical features of these plants. The key will not "work" for species 
which are in one way or another anomalous, e.g., for species of 
Myrcia with 4 calyx-lobes. It will suffice, however, to indicate gen- 
eric affinities for a great majority of all species in the Peruvian flora. 


1. Inflorescence myrcioid (i.e., compound, many-flowered, with the lower 
branches opposite and elongate; the upper branches progressively shorter 
and closer together; the uppermost often alternate because of abortion of one 
of a pair; the flowers tending to be aggregated toward the tips of the branch- 
lets and often solitary or in simple dichasia at the tips); calyx-lobes, if 
developed, usually 5 (Myrciinae). 

2. Calyx-lobes 5 (rarely 4), distinct even in the bud; principal axis of the panicle 
well developed, about as long as the primary lateral branches Myrcia. 

2. Calyx closed in bud, or only the tips of the lobes free, in anthesis circum- 
scissile or splitting irregularly; principal axis of the panicle often abortive 
at the lowest node and the panicle bifurcate. 

3. Calyx closed in bud, circumscissile, the calyptra often remaining attached 
at one side Calyptranthes. 

3. Calyx closed in bud, or the tips of the (usually 5) calyx-lobes free, the flower 
opening by irregular longitudinal slits between the lobes Marlierea. 

1. Inflorescence bifurcate (usually with sessile flowers in the forks), or racemose 
(the flowers or branches all in decussate pairs on an unbranched axis which 
aborts at tip), or the flowers solitary. 

4. Inflorescence racemose (species with glomerate or subumbellate flowers 
should be sought here; inspection will usually show that these flowers are in 
approximate, decussate pairs on a much shortened axis); calyx-lobes 4 and 
distinct, or rarely the buds closed or merely the tips of the lobes free (includes 
most of Eugeniinae). 

5. Hypanthium prolonged into a tube above the summit of the ovary; flowers 
mostly 4, sessile or nearly so, in conspicuously bracteate clusters; ovules 
2 in each locule. 

6. Calyx-lobes distinct, the tube circumscissile at base Myrciaria. 

6. Calyx splitting irregularly and longitudinally in anthesis, the bud completely 
closed or the tips of the calyx-lobes free Plinia. 

5. Hypanthium not prolonged above the ovary; calyx-lobes distinct; flowers 
few or many, rarely conspicuously bracteate (the bracteoles subtending in- 
dividual flowers often persistent and evident) ; ovules usually numerous. 


4. Inflorescence dichotomously branched, or the flowers solitary. 

7. Inflorescence dichotomously branched. 

8. Flowers 4-merous; calyx-lobes distinct and imbricate. 

9. Flowers often on old leafless branches, in regularly thrice-dichotomous cymes 
with a sessile flower in each fork; bracts persistent, in cuplike pairs; calyx- 
lobes persistent; flowers red, the stamens and style much longer (6-15 times) 
than the hypanthium Myrrhinium. 

9. Inflorescence axillary, on new leafy branches; dichasia usually 3- or 7- 

flowered, the forking branches variable in number and length; bracts usually 

deciduous; flowers white or yellowish, the stamens and style proportionately 

much shorter. 

10. Calyx-lobes deciduous Blepharocalyx. 

10. Calyx-lobes persistent Eugenia. 

8. Flowers 5-merous; calyx splitting irregularly in anthesis, the buds closed or 

the lobes short and widely separated Psidium. 

7. Flowers solitary. 

11. Calyx-lobes 4. 

12. Small-leaved shrubs (leaves less than 1 cm. long) of the high Andes; bracteoles 
foliaceous, elongate and persistent; anthers subrotund, the filaments filiform. 



12. Shrubs or trees with larger leaves, of middle and low elevations; bracteoles 
mostly deciduous, never persistent and foliaceous Eugenia. 

11. Calyx-lobes 5, or the calyx opening irregularly. 

13. Shrubs with small coriaceous leaves mostly less than 2 cm. long; high- 
montane Andean species; calyx-lobes 5, distinct; peduncles all 1-flowered; 
bracteoles foliaceous, elongate and persistent; anthers sagittate, the con- 
nective dilated Ugni. 

13. Shrubs of middle and low elevations, with larger leaves; flowers solitary or 
in dichasia, the bracteoles never foliaceous and persistent; filaments filiform, 
the anthers not sagittate. 

14. Calyx-lobes distinct, membranaceous, spreading in flower and in fruit; 
leaves with 3-4 pairs of the lowest lateral veins closely grouped and the upper 
increasingly distant Campomanesia. 

14. Calyx closed in the bud, or with short appressed lobes, the tube splitting in 
anthesis along irregular longitudinal lines; veins nearly uniformly spaced. 


One Peruvian genus, Acca, has purposely been omitted from the 
above key because its inclusion would complicate the key and make 
it more difficult of comprehension. One species of Acca has 
the flowers solitary or in simple dichasia; the other species has the 
flowers solitary or in congested racemes. The calyx-lobes are 4 in 
number in each species, and so are anomalous in the Pimentinae. 
The genus is indeed of uncertain systematic position but is easily 
recognized by the villous or tomentose foliage and the large red 

Notes on Individual Peruvian Genera 

The following conclusions have been reached after study of the 
criteria used by earlier workers, as set forth briefly above, and after 
the evaluation of these criteria in the light of new data, particularly 
from the inflorescence. Study of the vast myrtaceous flora of 
southeastern Brazil may demonstrate the wisdom of taking a some- 
what less provincial point of view with respect to the submergence 
of such genera as Abbevillea, Calyptromyrcia, Gomidesia, Rubachia, 
Siphoneugena, and others. The notes below are based primarily on 
materials from the Amazon basin and the northern Andes and should 
be considered with this in mind. 

1. The myrcioid genera. It seems clear that the Myrciinae are 
represented in Peru by three genera only. I cannot agree with Miss 
Amshoff that Aulomyrcia is a valid genus, even though it is possible 
to recognize certain species as members of the "genus." The 
characters of Aulomyrcia, which are readily observed in a series of 
small-flowered species typified by A. multiflora (Lam.) Berg, are 
two only: the hypanthium is prolonged above the summit of the 
ovary but is not constricted beneath the calyx; the disc and the 


inner surface of the hypanthium are usually glabrous. Among 
the Peruvian species, in my opinion, the distinction between Myrcia 
and Aulomyrcia breaks down completely; depending upon the species 
selected, one can demonstrate almost any desired stage of the tran- 
sition between the glabrous and prolonged hypanthium and calyx 
of Aulomyrcia on the one hand and the hairy ovary and short calyx 
of Myrcia proper. 

The status of Rubachia is not entirely clear and must be clarified 
by study of the several Brazilian species. The one Peruvian species 
referred by Berg to this genus is known in immature flowering 
condition only, and its relationships appear to be with other Peru- 
vian species of Myrcia. The genus Gomidesia, as far as I am aware, 
does not occur within or near the borders of Peru, although it has 
been reported from Bolivia; like Rubachia and Blepharocalyx, it is 
primarily a genus of southeastern Brazil. 

The distinction between Myrcia and Marlierea is a somewhat 
nebulous one, and it is probable that Marlierea comprises a phylo- 
genetically diverse assemblage of species which have been arbitrarily 
assigned to the genus because of the character of the irregularly 
splitting calyx. In spite of this doubt as to the monophyletic origin 
of Marlierea, its species are often recognizable as such by the per- 
sistent bracts and the abortion of the primary axis of the inflo- 
rescence. On the basis of the somewhat weak association of these 
characters of calyx, bracts and inflorescence, it seems best for the 
present to recognize the genus as distinct. 

Calyptranthes, like Marlierea, shows some evidence of internal 
diversity, but the calyx-character is such a distinctive one, and is 
correlated in so many species with the presence of dibranchiate 
hairs, that the genus is one of the most readily recognized among the 
American Myrtaceae. 

2. The eugenioid genera. The genera of the Eugeniinae are 
somewhat difficult to delimit, partly because of the paucity of 
material available for study and partly because of the many and 
morphologically diverse species. Whereas in the myrcioid groups 
the morphology of the inflorescence, the structure of the embryo, 
and the number of ovules in each locule remain essentially constant 
from genus to genus and even from species to species, in the eugen- 
ioid taxa quite the reverse is true. The number of species is large, 
and most of them have been assigned at one time or another to the 
genus Eugenia; any attempt at orderly classification of the whole 
group necessitates a decision as to which taxa shall be removed 


from the inclusive Eugenia, and as to the ultimate circumscription 
of Eugenia itself. 

The satisfactory partition of Eugenia on a worldwide basis has 
not yet been accomplished. It was thought at one time that the 
New World species could be separated from those of the Old World 
on the basis of floral characters. Merrill and Perry, 1 in their studies 
of Asiatic Myrtaceae, pointed out that such a separation was not 
practicable, but indicated that in their opinion the American species 
(Eugenia sens, str.) could be effectively distinguished from the 
Asiatic (the genus Syzygium Gaertn.) as follows: 

Eugenia: Embryo apparently undivided, pseudomonocotyledon- 
ous; seed-coat smooth and free from the pericarp; inflorescence 
centripetal, with the pedicels 1-flowered; calyx-limb less prolonged. 

Syzygium: Cotyledons distinct; seed-coat roughish, loosely or 
closely adhering to the pericarp; inflorescence usually centrifugal, 
with the panicles branching by threes or with secondary cymes; 
calyx-limb more prolonged. 

An opposing point of view has been taken more recently by 
Henderson 2 in his study of Eugenia (sens, lat.) in Malaya. In this 
author's opinion little dependence can be placed on the seed- 
characters to which Merrill and Perry attached primary importance. 
Henderson maintains Eugenia as a single worldwide genus to include 
not only Syzygium (which contains the great majority of the Old 
World species) but also Acmena DC. and Cleistocalyx Blume, both 
of which were recognized by Merrill and Perry as independent 
genera. Henderson also suggests tentatively that "if the Old and 
New World species are to be separated, better characters might be 
found in the structure of the inflorescences and flowers." 

Ingle and Dadswell, 3 in a paper on the wood anatomy of the 
Myrtaceae, support on anatomical grounds the view that at least two 
genera are probably represented in the inclusive Eugenia. A taxon 
which these authors call "Eugenia B" (that is, including Acmena, 
Cleistocalyx, Syzygium, and some Asiatic and Pacific species still 
referred to Eugenia) is shown to differ fundamentally from "Eugenia 
A" (that is, the New World species, plus the few Old World species 
referred to the genus Jossinia). Ingle and Dadswell examined 

Arnold Arb. 19: 99-100, 205-208. 1938; Mem. Amer. Acad. Arts, 
Sci. 18, pt. 3 [Mem. Gray Herb. 4]: 135-140. 1939. 

2 Gardens' Bull. Singapore 12, pt. 1: 1-17. 1949. 

3 The anatomy of the timbers of the south-west Pacific area. III. Myrtaceae. 
Austral. Journ. Bot. 1: 353-401. pis. 1-10. 1953. 


wood-samples of seven species of Eugenia thought to be tropical 
American in origin, and numerous samples of various species from 
the southwest Pacific. Although these authors do not support 
Merrill and Perry in the recognition of the small genera Acmena, 
Cleistocalyx and Jossinia, all are in general agreement that there 
is a notable hiatus between two great groups of species, one primarily 
of the Old World and the other of the New, that have been in the 
past indiscriminately assigned to Eugenia. 

The separation of "Eugenia A" and "Eugenia B" has received 
additional, if somewhat inconclusive, support from studies in the 
comparative morphology of pollen, by Kathleen M. Pike. 1 Miss 
Pike found the pollen of three species of "Eugenia A" (none of which 
was American) to be quite distinct from that of "Eugenia B." She 
was unable to distinguish the pollen of Cleistocalyx from that of 
Syzygium, but could distinguish these genera from Acmena on the 
basis of pollen morphology. 

An increasing body of evidence thus points to the conclusion that 
the New World eugenioid species are generically distinct from most 
of the Old World species formerly referred to Eugenia. It is probable 
that the name Syzygium Gaertn. will be conserved for the principal 
group of these Old World species, and it may be that Syzygium will 
attain general recognition as an independent genus. The propriety 
of such a course, as pointed out by Henderson, cannot be determined 
with much certainty unless someone is willing to undertake the 
thankless and difficult task of a worldwide revision of Eugenia. 
In the meantime it seems clear that the American species can 
profitably be studied apart from their Old World relatives, since as 
far as I am aware no one has suggested any close relationships at 
the specific level. 

Within the large group of American eugenioid species are several 
distinctive sub-groups which some authors have recognized as 

(a) Eugenia sens. str. This is the group which includes the type 
of the genus, E. uniflora L. In the approximately 70 species recog- 
nized by Berg (as members of his genera Phyllocalyx and Steno- 
calyx), the hypanthium is not prolonged beyond the ovary, the 
ovules are numerous, and the embryo is undivided; the calyx-lobes 
are often elongated and foliaceous or subfoliaceous, and the flowers 
occur partly in short racemes and partly in bracteate axils at the 

1 Pollen morphology of Myrtaceae from the south-west Pacific area, Austral. 
Journ. Bot. 4: 13-53. pi. 1. 1956. 


bases of new shoots. Berg based the genera Phyllocalyx and Steno- 
calyx on the characters of calyx-lobes and flower arrangement. It 
does not seem desirable to recognize these segregate genera (aside 
from the fact that one of them, Stenocalyx, was illegitimate when 
published for it included the type species of Eugenia), because the 
inflorescence character on which they are primarily based is one 
which occurs occasionally in racemose-flowered species of Eugenia. 
It is possible to demonstrate various intermediate conditions be- 
tween the one extreme, in which all flowers are in the axils of bracts 
at the bases of new shoots, and the other, in which the flowers are 
normally in short racemes (as in Eugenia egensis DC.). Phyllocalyx 
and Stenocalyx together, however, form a somewhat doubtfully 
recognizable taxon, already designated by Niedenzu (Natiirl. 
Pflanzenfam. Ill, pt. 7: 81. 1893) as Eugenia, subgenus Macrocalyx. 
The range of the group is from the West Indies southward through- 
out the lowlands of eastern South America. As yet no species has 
been found in Peru. 

(b) Typical Eugenia in the sense of most authors, not of Lin- 
naeus. A large group of species with multiovulate ovary, the 
hypanthium little or not at all prolonged beyond the ovary, and 
the flowers in short or long racemes or apparently glomerate because 
of extreme reduction of the raceme axis. These include most of the 
Biflorae, Corymbiflorae, Glomeratae, Racemosae, Racemulosae and 
Umbellatae of Berg, a total according to him of about 360 species 
in America. The embryo, as far as known, is undivided, as in typical 
Eugenia; most species are known from flowering material only, 
however, and it is not possible to make a general statement about 
embryo structure in this group. 

(c) The genus Anamomis Griseb. A group of 40 to 50 species, 
ranging from the West Indies southward, chiefly along the Andes, 
to Bolivia and Argentina. The ovary is multiovulate, the hy- 
panthium is little or not at all prolonged, the flowers are solitary or 
in simple or compound dichasia, and the cotyledons as far as known 
are 2 and distinct. The species of this taxon were included by Berg 
in his group Dichotomae, of Eugenia. On the basis of geographical 
restriction, the structure of the inflorescence, and the morphology 
of the embryo, Anamomis seems to be more deserving of independent 
generic status than Plinia, Myrciaria or Catinga. A revision is 
needed, however, of the entire group of taxa including Anamomis, 
the south-Brazilian Myrcianthes Berg, and the Chilean genera 
Reichea Kausel and Myrceugenella Kausel. Myrceugenella, for 


example, differs from Anamomis chiefly in having the radicle about 
as long as (instead of about one-third as long as) the cotyledons. 
Reichea and Myrcianthes have pentamerous rather than tetramerous 
flowers, but otherwise seem to differ little from Anamomis. At 
present, moreover, most of the species of Anamomis are unknown 
in the fruiting condition (mature fruits have been seen for 5 of the 
19 known species which occur in Peru), and in view of the importance 
which has been attached to the myrtaceous embryo as an indicator 
of relationship, it seems unwise to transfer numerous species to this 
genus, on the basis of the inflorescence alone, before the taxonomy 
of the group has been carefully studied. 

The nomenclature of the group is somewhat involved. If study 
indicates that but a single genus should be recognized, the oldest 
name for this is apparently Lama A. Gray, published in 1854 and 
thus antedating any of the generic names published by Berg. If 
Myrcianthes and Anamomis prove to be congeneric, the former name 
has priority. If Anamomis proves to be distinct from all the others, 
the name is valid for the Andean-West Indian group of species 
which is considered here. 

There appears to be little more than superficial similarity be- 
tween the species of Anamomis and those of the Old World genus 
Syzygium Gaertn. The inflorescence in Syzygium is often terminal; 
if lateral it may be strictly racemose (with the terminal flower 
developed, not abortive as in most American Eugenias), or more 
often it may simulate the myrcioid panicle, with the principal 
branches opposite, elongate, and terminating in single flowers or 
small dichasia. The consistent occurrence of a regularly dicho- 
tomous dichasium is apparently unique in Anamomis and its rela- 
tives (cf. similar inflorescences in Pimentinae). 1 

1 Since the above was set in type I have received, through the kindness of 
Dr. Carl Skottsberg, a copy of a very recent paper in which the whole matter 
of generic segregations in the American Myrtaceae is discussed by Dr. Eberhard 
Kausel (Beitrag zur Systematik der Myrtaceen, Ark. Bot. 3: 491-516. 1956). 
Kausel proposes to erect a new family, Leptospermaceae, for the capsular-fruited 
genera usually assigned to Myrtaceae, and to restrict the use of the name Myr- 
taceae to those fleshy-fruited genera which comprise the tribe Myrteae DC. 
The Myrtaceae in this narrower sense are divided by Kausel, on the basis of 
characters of the seed and embryo, into the following taxa, which are presumably 
intended as subfamilies, although apparently not formally designated as such: 
Eugenioideae, Plinioideae, Cryptorhizoideae, Myrtoideae, and Myrcioideae. 
Five new genera are proposed, of which one is assigned to Eugenioideae, three to 
Plinioideae, and one to Myrtoideae. In the Plinioideae Kausel recognizes Ana- 
momis Griseb. and four other genera which have the inflorescence an axillary 
dichasium, viz., Pseudanamomis Kausel, Reichea Kausel, Myrcianthes Berg, 
and Amyrsia Raf. with type species Myrtus foliosa H.B.K. The superficially 
similar Pseudomyrcianthes Kausel is assigned to the Eugenioideae and Myrceu- 
genella Kausel to the Myrcioideae. 


(d) The genus Myrciaria Berg. A group probably of about 
40 species with few (usually 2) ovules in each locule of the ovary, 
the hypanthium markedly prolonged beyond the ovary and circum- 
scissile at the base, the flowers subsessile in small (racemose) clusters 
and the cotyledons distinct or nearly so. According to Berg the 
genus included 65 species; Bentham pointed out (Benth. & Hook, 
f. Gen. PI. 1: 720. 1865) that several of the original species probably 
belonged to some myrcioid genus. After the removal of these 
anomalous species the genus Myrciaria appears to comprise a co- 
herent and natural group, which ranges from Panama southward, 
chiefly in the lowlands east of the Andes. Siphoneugena Berg is 
a very small south-Brazilian genus of uncertain status which differs 
from Myrciaria chiefly in having the ovary multiovulate. 

(e) The genus Plinia L., in the sense of Urban (Repert. Sp. 
Nov. 15: 412-413. 1919). A small assemblage of species in which 
the locules of the ovary are 2-ovulate, the hypanthium is prolonged, 
the calyx is partially or completely closed in bud and splits longi- 
tudinally at anthesis, the flowers are usually in sessile bracteate 
clusters, and the cotyledons are distinct. About half a dozen closely 
related species are known from northern continental South America; 
Urban described in his later years a considerable number of species 
of Plinia from the West Indies, but most of these are small-leaved 
plants with little apparent similarity to the South American species. 
On the basis of the distinctive calyx and the small number of ovules, 
the South American species are readily recognized as members of 
this group, and I have assigned several Peruvian species to it. 

(f) A heterogeneous group of perhaps 15-20 species, in which 
the ovary is multiovulate, the hypanthium is mostly not or scarcely 
prolonged beyond the ovary, the buds are closed or the calyx-lobes 
coherent in the bud, the flowers are in short axillary racemes, and 
the embryo is undivided. Except for the united or coherent calyx- 
lobes, any of these species could pass for a member of the genus 
Eugenia, and many of them have been attributed at times to that 
genus. The taxon characterized above includes Calycorectes Berg 
(with buds closed) and probably also Schizocalyx Berg, and in Peru 
and Amazonian Brazil the genus Catinga Aubl. (in the sense of 
Amshoff, in Fl. Suriname 3: 105. 1951). The status of Catinga is not 
clear to me, and I hesitate to recognize it as an independent genus. 
It differs from Eugenia in one character only, namely, that the four 
calyx-lobes are coherent in the bud for part or all of their length and 


are usually thickened and cucullate at the tips. 1 The species 
assigned to Catinga presumably differ from Calycorectes because in 
the former the lobes are loosely coherent along narrow lines near 
the edges but separate cleanly at anthesis into symmetrical thin- 
edged lobes, instead of rupturing irregularly as in Calycorectes. 
To me this is not a convincing character upon which to base a genus, 
especially since the supposed species of Catinga are not otherwise 
very different from the many species of Eugenia which have similar 
foliage and similar inflorescences. The genus Calycorectes itself is 
not well understood; it is indeed restricted by Amshoff (op. cit. 104) 
to a single species, although more than 15 species have been de- 

In summation, the Peruvian eugenioid species belong for the 
most part to Eugenia, which is understood to include Phyllocalyx 
Berg and Stenocalyx Berg, and, for the present at least, Anamomis 
Griseb. and Catinga Aubl. Myrciaria Berg is apparently a distinct 
genus, and Plinia L. (in the sense of Urban) is recognized. 

3. The pimentoid genera. The genera of Pimentinae which 
have membranaceous seed-coats, namely Blepharocalyx and Cam- 
pomanesia, are represented in Peru by a single species each; evalu- 
ation of their generic limits should be undertaken only after study 
of the species of eastern and southeastern Brazil, where these genera 
are represented by additional species. The remaining genera, those 
with hard or bony seeds, fall into two principal groups, namely, 
Psidium and Myrtus and their respective segregates. The separation 
between these two has been traditionally upon the basis of calyx 
morphology. Myrtus has been a form-genus including especially all 
small-leaved species (like the type species, M. communis L., of 
Europe) with 1-flowered peduncles, 2- to 3-locular ovary and 5 
(rarely 4) distinct calyx-lobes. Psidium was originally a form-genus 
including mostly larger-leaved species (like the type species, P. 
guajava L., the guava) with 1- or 3-flowered peduncles, the ovary 
usually more than 2-locular, and the calyx closed or nearly closed 
in bud and splitting irregularly in anthesis. 

In tropical America Psidium is a relatively "good" genus, with 
numerous and distinctive species, but unfortunately the character 
of the closed and irregularly breaking calyx applies to relatively few 
of these. Many species described by DeCandolle, Berg, and later 
authors have an open, lobed calyx. The lobes are usually small, and 

1 For additional notes on the status of Catinga, see below (p. 203), in the 
discussion of Eugenia acrensis. 


the hypanthium splits longitudinally in the sinuses as the flowering 
period passes and the fruit expands; this constitutes, in fact, one of 
the principal ways in which a species of Psidium may be recognized. 
Calyptropsidium Berg and Psidiopsis Berg differ in minor features 
of the calyx and are not convincingly distinct from Psidium. 

Seemingly intermediate between Psidium and Myrlus is Caly- 
colpus, which according to Riley 1 contains 12 species. This is 
a genus which has almost no characters. According to Riley it may 
be distinguished from Campomanesia, Myrtus and Psidium by its 
calyx-lobes, "which are patent in the flower-bud, whereas in Cam- 
pomanesia and Myrtus they are appressed to the corolla when in 
bud, and in Psidium they form a closed calyx which splits into 
segments as the flower expands." It is true that a species of Caly- 
colpus can usually be recognized as such by the rather large flowers 
which are solitary in the axils, by the calyx-lobes which are disposed 
as noted above, and by the coriaceous leaves, which usually blacken 
in drying. Superficially, however, a small Calycolpus looks not 
unlike a large specimen of Myrtus communis, and if any revision of 
Myrtus is undertaken the status of Calycolpus should be considered 
at the same time. As far as I am aware, no species of this genus 
ranges as far southwest as the borders of Peru; the Rio Negro forms 
the approximate limit of the group in this direction. 

The genus Myrtus itself is in need of study. Bentham (Benth. 
& Hook. f. Gen. PL 1: 714. 1865) stated that the genus as he knew it 
contained probably about 50 species, mostly American but occurring 
on all major land areas except possibly tropical Asia. Actually 
Myrtus in the sense of Bentham was something of a catch-all. It 
included the small Andean groups now generally referred to the 
genera Myrteola Berg and Ugni Turcz. ; Anamomis and Myrcianthes 
of the Eugeniinae; Myrceugenia of the Myrciinae; the Bergian 
genera Blepharocalyx and Pseudocaryophyllus of the Pimentinae; 
additional Chilean genera now regarded by Kausel as completely 
distinct; and various additional American plants, practically all of 
which have since been referred to genera other than Myrtus. 

At least one recent author has retained a few American species 
in Myrtus, 2 but an opposing point of view has been taken by Burret, 3 
who has transferred most of the described species (from both Old 

1 Kew Bull. 1926: 145-154. 1926. 

2 Legrand, D. Las Mirtaceas del Uruguay, An. Mus. Hist. Nat. Montevideo, 
ser. 2, 4, pt. 11: 21-24. 1936. 

3 Myrtaceen-Studien, Notizbl. Bot. Gart. Berlin 15: 479-500. 1941. 


and New Worlds) to other genera, retaining in Myrtus no more than 
the original M. communis, one African species and about 14 species 
in Florida and the Greater Antilles. 

Pending further revision of the whole group as it exists in 
America, I have chosen to recognize Ugni and Myrteola as local, 
independent genera. These appear well founded when judged by 
ordinary criteria of morphological distinctness, homogeneity, and 
well-defined geographical range. It is by no means clear that Myrtus 
is represented in the American flora by any species whatsoever, and 
it seems unwise to attempt to force the several Myrtus-\ike American 
groups into a generic concept based so largely upon a single European 

Finally, a word should be said about Acca, which Berg placed 
among the Eugeniinae. Bentham stated his opinion that the genus 
belonged rather with Psidium, and the seed-characters of the two 
species of Acca confirm this disposition. The genus is well marked, 
with style and androecium unlike anything known in the Myrteae 
unless it be Feijoa Berg, which Burret (Repert. Sp. Nov. 50: 49. 
1941) has relegated to the synonymy of Acca. 

Systematic Treatment of Genera and Species 
1. MARLIEREA Camb. 1 

1. Midvein flat or convex on the upper surface of the leaf, up to 1-1.5 mm. 
broad at base, if sulcate this at base of blade only. 

2. Inflorescence, including the hypanthium, strongly velutinous or at least with 
numerous loosely ascending reddish or reddish-yellow silky hairs up to 
1 mm. long. 

3. Inflorescence strongly velutinous; flowers large, the buds 7 mm. long; calyx- 
tips free; leaves acute or acuminate, 13-18 cm. long. . .M. velutina McVaugh. 

3. Inflorescence loosely silky-hairy; flowers small, the buds 2.5 mm. long, 
closed, apiculate; leaves caudate-acuminate, 4.5-7 cm. long. 

M. caudata McVaugh. 

2. Inflorescence glabrous or essentially so; at least the tips of the calyx-lobes 
evident in the bud. 

4. Leaves 9-14 cm. long, 2-3.5 times as long as wide; lateral veins 12-15 pairs, 
the transverse veins obscurely reticulate; leaves finely and obscurely dark- 
dotted; calyx-lobes in bud minute, ciliate M. scytophylla Diels. 

4. Leaves 7.5 cm. long or less, 1.7-2.2 times as long as wide; lateral veins 6-8, 
the transverse veins prominently and coarsely reticulate; leaves with 1-3 
large translucent dots per square millimeter; calyx-lobes in bud distinct, the 
inner ones broadly scarious-margined, 2.5 mm. wide. .M. areolala McVaugh. 

1 One species, Marlierea insignis, is described below but is not included in the 
key; it is known from Amazonian Colombia (Rio Apaporis) and is probably not 
a member of the Peruvian flora. 


1. Midvein sharply and narrowly impressed on the upper surface, or in one 
species concave or broadly sulcate. 

5. Inflorescence, including buds, with numerous ascending lustrous yellowish- 
white hairs up to 1 mm. long; buds 4-5 mm. long, mostly concealed by the 
hairs; mid vein concave or sulcate; hypanthium tomentose within. 

M. spruceana Berg. 

5. Inflorescence pubescent, often sparsely so, with short, pale or reddish hairs 
0.5 mm. long or less; buds glabrous or essentially so, 1.5-3.5 mm. long; 
mid vein sharply and narrowly impressed; hypanthium glabrous within. 

6. Leaves cordate-auriculate, nearly sessile, the petioles 3-4 mm. long; in- 
florescence finely hispidulous with minute stiff erect hairs; buds closed, with 
a prominent narrow apiculum; staminal ring short-hairy. 

M. subulata McVaugh. 

6. Leaves acute to cuneate or somewhat rounded at base, on petioles 4-10 mm. 
long; inflorescence pubescent with appressed or ascending hairs; buds closed 
or the calyx-lobes distinct; staminal ring glabrous. 

7. Petioles transversely rimose, the reddish-brown or whitish papery layers 
separating but persistent; calyx in bud with 4 very small separate deltoid 
tips; lower branches of the panicle straight and elongated, spikelike, with 
numerous sessile flowers and short squarrose bracts. 

M. umbraticola (HBK.) Berg. 

7. Petioles smooth and with unbroken surface, usually dark in color; buds 
closed, or with 5 distinct calyx-lobes; inflorescence various. 

8. Inflorescence thinly pubescent with pale hairs; bracts and bracteoles decidu- 
ous before anthesis; lateral veins of the leaves not impressed above; buds 
3-3.5 mm. long, closed at apex M. imperfecta McVaugh. 

8. Inflorescence pubescent with lustrous rufous hairs; bracts and bracteoles 
persistent, squarrose; lateral veins impressed above; buds 1.5-2 mm. long, 
with 5 distinct calyx-lobes M. squarrosa McVaugh. 

Marlierea areolata McVaugh, sp. nov. 

Subglabra, multiflora, foliis 7.5 cm. longis; venis utroque latere 6-8, venulis 
reticulatis; laminis utrinque crebro pellucido-punctatis; nervo medio supra piano 
vel convexo; calycis lobis liberis 4, hypanthio demum explanato, ad germinis 
verticem partito. 

A shrub or tree with elliptic, bluntly acuminate leaves and relatively few 
lateral veins, the foliage distinctive also because of the prominently reticulate 
veinlets and the numerous large translucent glands; calyx-lobes in unequal pairs, 
the inner ones broadly scarious-margined, about 1.7 mm. long and 2.5 mm. wide; 
style 4 mm. long; stamens 75-100. 

The large and distinct calyx-lobes, and the deciduous bracts and 
bracteoles, suggest that this species may have some affinity to the 
genus Myrcia (Aulomyrcia) . It is here referred, however, to Mar- 
lierea, because of the flattening of the hypanthium after anthesis, 
and the accompanying distortion and splitting of the calyx. As an 
additional minor character may be mentioned the terminal or 
falsely terminal inflorescence in this species; this character recurs 
throughout the genera Calyptranthes and Marlierea, whereas in 
Myrcia the panicles are more often from the lower axils. 


Peru, Loreto: Stromgebiet des Ucayali von 10 S. bis zur Miin- 
dung, G. Tessmann 3264, anno 1923 (G, type). This specimen was 
determined in the herbarium by Burret as a new species of Myrcia, 
but as far as I can learn this has never been published. F.M. Neg. 

Marlierea caudata McVaugh, sp. nov. 

Arbor, ramulis paniculisque ferrugineo-sericeis; foliis parvis, usque ad 7 cm. 
longis, longe acuminatis; paniculis paucifloris, alabastra 2.5 mm. longa, clausa, 

A tree up to 7 meters high, the pubescence of reddish or reddish-yellow silky 
hairs; leaves 4.5-7 cm. long, 2-3 times as long as wide (including the narrow 
acuminate tip 1-2 cm. long); principal branches of the inflorescence 3-5 cm. long, 
each with 20 flowers or fewer; calyx splitting irregularly into 4 oblong lobes; style 
4.5 mm. long; stamens about 100; fruit subglobose, 8-11 mm. in diameter. 

Peru, Loreto: Mishuyacu, near Iquitos, elev. 100 meters, forest, 
G. King 235, Oct.-Nov., 1929 (F 624286, type; NY; US); forest 
between [lower] Rio Nanay and Rio Napo, June 6, 1929, L. Williams 
718 (F). 

Marlierea imperfecta McVaugh, sp. nov. 

Arbor, pubescens, foliis 12-25 cm. longis acuminatis, petiolis 5-8 mm. longis 
canaliculatis; nervo medio supra impresso, venis lateralibus non impressis; panic- 
ulis multifloris, bracteis bracteolisque deciduis; alabastris clausis, 3-3.5 mm. 

A small tree, thinly pubescent with pale or reddish-based hairs; leaves elliptic, 
2.5 times as long as wide, with 10-15 pairs of lateral veins; inflorescence 3 times 
compound, the principal branches 8-14 cm. long; calyx splitting irregularly into 
4 lobes; style 4-4.5 mm. long; stamens about 100. 

Peru, Loreto: Mishuyacu, near Iquitos, elev. 100 meters, forest, 
Jan., 1930, G. Klug 787 (US 1455778, type; NY; F). 

Marlierea insignis McVaugh, sp. nov. 

Arbor mediocris, insigniter velutina; ramulis, petiolis paniculisque, pilis 
rigidis erectis usque ad 1 mm. longis, aureo-fulvis, dense obtectis; foliis ellipticis 
15-25 cm. longis acuminatis; venis utroque latere 12-15, planiusculis vel impressis; 
calyce in alabastro ut videtur clauso, ad florendi tempus irregulariter rumpente; 
paniculae ramis 4-6 cm. longis crassis, paucifloris; fructu globoso, diametro 
1.5-2 cm. 

Tree up to 10 meters high, densely velutinous with golden-brown hairs, the 
hairs stiff, erect, up to 1 mm. long, completely covering the branchlets, vegetative 
buds, inflorescence (including the fruit), and mid vein at least beneath; lower 
leaf-surface loosely hirsute with similar hairs; leaves broadly elliptic, 7-13 cm. 
wide, 15-25 cm. long, rounded to apex and then abruptly short-acuminate, the 


acumen acute with concave sides, 5-15 mm. long; base of blade rounded, the 
margins at very base decurrent on the petiole, which is up to 4 mm. thick (including 
hairs) and 10 mm. long; midvein flat or concave above, velutinous or glabrate, 
prominent beneath; lateral veins 12-15 pairs, plane or impressed above, prominent 
beneath; marginal vein about equaling the laterals, strong and little arched 
between them, 3-6 mm. from margin; leaves dull and pale green above, yellow 
green beneath and there obscurely and minutely dark-dotted; inflorescence 
branches 3-6 cm. long, 3 mm. thick (including the hairs), solitary in the axils 
or the fertile branch aborting at the first node and producing a pair of nearly 
equal lateral branches; individual branches spike-like, with up to 7 nearly sessile 
flowers; flowers not seen; fruit globose or nearly so, about 1.5 cm. long, 1.3-2 
cm. in diameter; calyx apparently closed in bud, with a conical top, glabrous 
within this, and breaking irregularly in anthesis; inner disc at base of style 2 mm. 
in diameter, the hypanthium produced 1-1.5 mm. above the summit of the ovary 
and hairy within, the numerous (probably 150-200) stamens produced near the 
summit of the bud in a ring 1.5 mm. wide; anthers 0.3 mm. long; style 5 mm. long; 
ovules 2 in each of 2 locules. 

Colombia, Amazonas-Vaupe"s: Rio Apaporis, Soratama, elev. ca. 
250 meters, June 21, 1951, R. E. Schultes & I. Cabrera 12766 (US 
2220050, type; MICH) ; Aug. 24, 1951, Schultes & Cabrera 13722 (US). 

Marlierea squarrosa McVaugh, sp. nov. 

Frutex, rufo-pubescens; foliis 9-13 cm. longis acuminatis, venis supra im- 
pressis, petiolis 5-7 mm. longis levibus; floribus subspicatis, bracteis divaricatis 
persistentibus; alabastris 1.5-2 mm. longis glabris; calycis lobis 5 valde inaequal- 

A shrub 2 meters high, with elliptic leaves 2.5 times as long as wide, and 
many-flowered panicles; flowers small, the calyx splitting irregularly from the 
base of the lobes to the summit of the ovary; style 3.5 mm. long; stamens about 50. 

This species, like Marlierea areolata, seems to cross the supposed 
generic lines between Marlierea and Myrcia (Aulomyrcia) . It has 
the irregularly splitting calyx, persistent bracts and explanate disk 
of Marlierea, but the coppery color (when dry) and free calyx-lobes 
of some species of Myrcia. It might conceivably be a hybrid in- 
volving Marlierea umbraticola, which also has impressed veins, but 
as yet M . umbraticola has not been collected within the borders of 
Peru. The following is the only known collection of M. squarrosa. 

Peru, Loreto: Mishuyacu, near Iquitos, elev. 100 meters, forest, 
G. King 169, Oct.-Nov., 1929 (US 1455168, type; NY; F). 

Marlierea subulata McVaugh, sp. nov. 

Arbor, hispidula; foliis subsessilibus, 13-18 cm. longis, ovatis lanceolatisve ; 
nervo medio supra impresso; paniculis angustis paucifloris; alabastris subglabris, 
3 mm. longis, clausis, apiculatis; disco staminali pubescente. 


A tree 5 meters high, with cordate-auriculate nearly sessile leaves about 3 
times as long as wide; bracts subulate, up to 3 mm. long, more or less persistent 
through anthesis; calyx splitting irregularly into 4 lobes; style 6 mm. long; stamens 
75-100; fruit up to 1 cm. long and 1.5 cm. in diameter. 

Peru, Loreto: Mishuyacu, near Iquitos, elev. 100 meters, forest, 
May, 1930, G. King 1341 (F 627613, type; NY; US). 

Marlierea velutina McVaugh, sp. nov. 

Fulvo-brunneo-velutina, foliis 13-18 cm. longis acuminatis, nervo medio 
supra piano vel basi sulcato; paniculis multifloris; floribus grandis, alabastris ut 
videtur 7 mm. longis; hypanthio circiter 8-sulcato; calyce irregulariter rumpente. 

A shrub or tree, densely beset with coarse yellowish-brown hairs up to 1 mm. 
long; leaves elliptic to lanceolate, about 3 times as long as wide; inflorescence 
6-10 cm. long, 2 or 3 times compound; calyx-lobes 4, at first united except the free 
tips 1 mm. long; style 7 mm. long; stamens probably about 200. 

A distinctive species suggesting some of the members of the 
genus Gomidesia because of the color and character of the pu- 

Brazil, Guapore": Falls of Madeira, H. H. Rusby 2683, Oct., 1886 
(F, type; US). Univ. of Mich. Neg. 449. 


1. Flowers very large for the genus, the buds 7-8 mm. long, convex or nearly 
flat at the apex, lacking a narrow apiculum; inflorescence pale-scurfy and 
also appressed-puberulent with minute brownish dibranchiate hairs 0.1 mm. 
long; leaves 25-39 cm. long, narrowly elliptic, appearing cordate-auriculate 
at base, nearly sessile, the stout petiole 4 mm. long. . .C. maxima McVaugh. 

1. Flowers smaller, the buds if 5 mm. long or more fusiform and apiculate, or 
noticeably hirsute or velutinous; inflorescence glabrous to hairy; leaves 
usually smaller and slender-petiolate, if sessile or essentially so the buds not 
as above. 

2. Leaves sessile, with veins impressed on the upper surface; blades cordate 
at base, or the margins much produced and puckered. 

3. Leaves ovate, cordate, 9-12 cm. long; inflorescence glabrous, the buds 
fusiform, 6-7 mm. long C. sessilis McVaugh. 

3. Leaves obovate, with the margins near base produced into puckered folds, 
the blades 30-38 cm. long; inflorescence appressed-hirsutulous, the buds 
obovoid, 3.5 mm. long C. plicata McVaugh. 

2. Leaves petiolate, cuneate to acute or somewhat rounded at base, the veins 
usually not impressed above. 

4. Inflorescence of paired spikes, the individual flowers sessile along the axis, 
or the lowest in sessile or very short-peduncled groups of three. 

5. Buds glabrous; leaves 2.5-6 cm. long, often obovate with rounded or some- 
times short-acuminate tips; flowers 3-5 (-11) in each spike. . C. pulchella DC. 

5. Buds strigose or hirsute; leaves 7-16 cm. long, elliptic or ovate, usually 
prominently and often narrowly acuminate; flowers more numerous. 


6. Branchlets and inflorescence, including the buds, thickly rufous-hirsute; buds 
broadly obovoid to nearly globose, 5-6 mm. long, concealed by the hairs; 
flowers 8-13 in each spike C. krugioides McVaugh. 

6. Branchlets and inflorescence with appressed yellow or brown dibranchiate 
hairs; buds obovoid or broadly fusiform, 2-2.5 mm. long, rather sparingly 

7. Spikes mostly 8-12 cm. long, the numerous flowers in several sessile clusters 
of 10-20 flowers each; leaves short-acuminate, the lateral veins not im- 
pressed above, slender and closely parallel C. densiflora Berg. 

7. Spikes 3-5.5 cm. long, the flowers 25 or fewer in each, in small sessile clusters 
of 1-3 each; leaves with narrow acumen 1.5-2 cm. long; lateral veins im- 
pressed above, the principal ones prominent beneath and contrasting with 
the less conspicuous intermediates C. brevispicata McVaugh. 

4. Inflorescence of paired panicles or compound dichasia, the basal branches 
elongate and again branched, or occasionally 1-flowered only. 

8. Branches of the inflorescence variously pubescent (sometimes thinly so) 
with appressed or erect hairs; hypanthium variously strigose to tomentose 
except in C. multiflora. 

9. Hypanthium glabrous; inflorescence loosely pubescent with numerous erect 
or somewhat appressed soft pale rufous hairs; midvein convex above; buds 
2-2.5 mm. long C. multiflora Berg. 

9. Hypanthium strigose or variously velutinous or tomentose, if nearly glabrous 

the midvein sulcate or narrowly impressed above. 

10. Inflorescence a compact umbelliform cyme 2-5 cm. long with 15 flowers or 
fewer, the ovate boat-shaped bracts sub-foliaceous, persistent; inflorescence, 
including the flowers, thickly hirsute with coarse sessile dibranchiate hairs 
up to 1.5-2 mm. long and attached near one end C. longifolia Berg. 

10. Inflorescence paniculate or by reduction racemose, often 3 to 4 times com- 
pound and many-flowered (if short and few-flowered not hirsute as above); 
bracts all deciduous before anthesis, or a few (usually the basal ones) per- 
sisting; hairs of the inflorescence various, mostly sessile and less than 1 mm. 

11. Hairs of the inflorescence golden-yellow, dibranchiate, up to more than 1 mm. 
long, the basal stalk of the hair erect and often as long as the spreading or 
ascending branches; leaves 6 cm. long or less, rounded to obscurely acuminate 
at tip; branchlets 2-winged; flowers mostly 10 or fewer on each branch. 

C. tridymantha Diels. 

11. Hairs of the inflorescence sessile and, if abundant, usually red or rusty, all 
somewhat appressed and mostly less than 0.5 mm. long, or with very short 
hairs intermixed, or the inflorescence velutinous or tomentose. 

12. Leaves large, mostly more than 15 cm. long (often 20-30 cm.), with 20-35 
pairs of lateral veins; inflorescence with abundant rufous pubescence; buds 
obovate or obconic, scarcely apiculate, 3-5 mm. long. 

13. Blades tapering from the middle or below, to a slender apex; lateral and 
marginal veins scarcely apparent on the lower leaf surface, which is covered 
with very numerous closely appressed pale hairs up to 0.2 mm. long; inflo- 
rescence 5 cm. long or less, few-flowered C. macrophylla Berg. 

13. Blades abruptly and narrowly acuminate; lateral and marginal veins forming 
a conspicuous pattern on the lower surface, which is glabrous or sparingly 
appressed-pubescent; inflorescence 6-10 cm. long, many-flowered. 

C. gigantifolia McVaugh. 

12. Leaves of moderate size, usually less than 20 cm. long or, if longer, the 
lateral veins 15 pairs or fewer, or the buds 2-2.5 mm. long; buds and pu- 
bescence various. 


14. Flowers small, the buds 2-2.5 mm. long, obovoid, the apex rounded or shortly 
apiculate; panicles mostly 3 times compound, many-flowered, the branches 
sparingly covered with appressed pale or sometimes reddish hairs; hypan- 
thium strigose, sometimes very sparingly so; paired panicles from an abortive 
flattened axis 1 cm. long or less. 

15. Lower leaf surfaces with few dark hairs and usually with rather numerous 
persistent nearly colorless appressed hairs; leaves elliptic, broadest at the 
middle, 5-10 cm. long, with 12-15 pairs of lateral veins; style 4-4.5 mm. long. 

C. ruiziana Berg. 

15. Lower leaf surfaces glabrous except for a few dark hairs; leaves ovate or 
lanceolate, usually widest somewhat below the middle, 9-15 (-25) cm. long, 
with 20-25 pairs of veins; style 5-6 mm. long C. simulata McVaugh. 

14. Flowers larger, the buds 3-6 mm. long, variously shaped; panicles compound, 
or by reduction racemoid, the branches and the hypanthium usually con- 
spicuously appressed-hairy, velutinous, or tomentose, with ferruginous or 
dark reddish hairs; panicles paired or with a central axis. 

16. Inflorescence a pair of spikes with all flowers sessile, or the lower branches 
1-2 cm. long and 1-flowered, or with 3 sessile flowers at the tip; buds 5-6 mm. 
long, abundantly hirsute, the hypanthium hairy within. 

C. krugioides McVaugh. 

16. Inflorescence paniculate, usually many-flowered and 3 times compound; 
buds 3-4 mm. long, appressed-hairy to tomentose, the hypanthium glabrous 

17. Mid vein impressed above; hypanthium appressed-hairy, the hairs of the 
inflorescence rusty-brown; buds fusiform, 3-3.5 mm. long; panicle narrow, 
the lowest branches about 1.5 cm. long C. tessmannii McVaugh. 

17. Midvein convex above (and then sometimes sulcate) or raised in a narrow 
ridge; hypanthium loosely velutinous or tomentose; buds obovoid or ellipsoid; 
lower branches of the panicle relatively long. 

18. Leaves 15-21 cm. long, about 4 times as long as wide, the straight marginal 
vein and the 10-15 short divaricate lateral veins prominent beneath; petiole 
very stout, 3 mm. thick and 7 mm. long; buds broadly ellipsoid, heavily 
tomentose with dark red hairs, the hypanthium urceolate in anthesis. 

C. rufotomentosa McVaugh. 

18. Leaves 10-16 cm. long, about 2.5 times as long as wide, the marginal vein 
not well defined, or consisting of a series of loops or arches between the 8-12 
pairs of arcuate laterals; petiole 1 mm. thick, 8-10 mm. long; buds obovoid, 
sparingly or rather densely velutinous with rusty brown or reddish hairs, the 
hypanthium probably turbinate in anthesis C. cuspidata DC. 

8. Branches of the inflorescence, and the hypanthium, completely glabrous or 
with a very few scattered hairs about the base and the nodes of the panicle. 

19. Panicles with 20 flowers or fewer, sometimes reduced and raceme-like or 
spike-like; peduncle and rachis filiform or very slender, often terete and 
nearly straight, usually less than 1 mm. thick; branchlets narrowly 2-winged. 

20. Flowers mostly sessile, only the lower clusters pedunculate; wings of the 
branchlets often 0.5 mm. high; buds 4-7 mm. long, obtuse or obscurely 
apiculate C. pulchella DC. 

20. Flowers mostly on very long slender pedicels; wings of the branchlets scarcely 
higher than thick; buds 2-3 mm. long, narrowly and conspicuously apiculate. 

C. bipennis Berg. 

19. Panicle many-flowered, 3 to 4 times compound, the peduncle usually some- 
what angular near summit and 1-1.5 mm. thick, the rachis often irregularly 
enlarged and zigzag; wings, if produced on branchlets, not persisting through 
a growing season. 


21. Lower leaf surface obscurely gland-dotted, the dots about 15 per square 
millimeter; leaves relatively narrow, mostly 2.5 times as long as broad, or 
longer; panicle branches irregularly alternate C. paniculata R. & P. 

21. Lower leaf surface prominently dark-dotted, the dots more than 50 per 
square millimeter; leaves broader, mostly 2.3 times as long as broad, or less; 
panicle branches, both large and small, often verticillate or fasciculate. 

C. crebra McVaugh. 

Calyptranthes brevispicata McVaugh, sp. nov. 

Arbor pubescens, foliis longe acuminatis, venis supra impressis, subtus 
prominulis, venulis interjectis tenuioribus; spicis 3-5.5 cm. longis binis, floribus 
usque ad 25 sessilibus, fasciculatim 1-3-nis e nodis oriundis. 

A small tree with elliptic-lanceolate leaves 11-15 cm. long, sparingly pu- 
bescent with yellowish brown hairs; a species superficially suggesting C. densiflora 
Berg, but the spikes shorter and with far fewer flowers than in that species; buds 
fusiform, about 2.5 mm. long; style 6 mm. long; stamens about 50. 

Peru, Loreto : Florida, Rio Putumayo, at mouth of Rio Zubineta, 
elev. 200 meters, forest, March-April, 1931, G. King 2040 (NY; US 
1456699, type). 

Calyptranthes crebra McVaugh, sp. nov. 

Arbor subglabra, floribus paniculatis, paniculae ramis ramulisque verticillatis 
vel fasciculatis; foliis ellipticis vel ovatis, 1.8-2.3-plo longioribus quam latioribus, 
subtus creberrime atro-punctatis. 

A tree 4-10 meters high, nearly glabrous, with leaves 8.5-10.5 cm. long; 
a species superficially like C. paniculata R. & P., but readily distinguished from 
that species by the key characters. Buds obovate, apiculate, about 2 mm. long; 
style 5 mm. long; stamens 60-75; fruit globose, about 5 mm. in diameter. 

Peru, Loreto: Near Iquitos, in forest, elev. 100 meters, Oct.-Nov., 
1929, G. King 77 (F; NY; US), 78 (F 624362, type; NY; US), 607 
(F; NY; US). 

Calyptranthes gigantifolia McVaugh, sp. nov. 

Arbor grandis, foliis 14-38 cm. longis acuminatis, subtus glabris vel parce 
pilis appressis pubescentibus, venis prominentibus; paniculis pilis appressis rufis 
vestitis, 6-10 cm. longis multifloris; alabastris 3-5 mm. longis. 

A tree up to 20 meters high, resembling C. macrophylla Berg, but that species 
having the leaves more coriaceous, acute rather than acuminate, and less promi- 
nently veined; the marginal vein in C. macrophylla is close to the margin and not 
impressed above; the lower surface in C. macrophylla is densely pubescent with 
minute appressed pale hairs and with numerous intermixed dibranchiate hairs; 
in this last respect C. macrophylla agrees precisely with C. speciosa Sagot, of the 
Guianas, and it may be identical with that species, as already suggested by 


Colombian specimens collected by Mutis (nos. 1940, 2234, 2951, 
3961, and 5754) appear to be conspecific with the Peruvian speci- 
mens cited below. 

Peru, San Martin: Juanjui, G. King 4277 (A; F; US 1458692, 
type). Loreto: Florida, Klug 2332 (F; G; GH; NY; US), 2347 
(F; G; GH; NY; US). 

Calyptranthes krugioides McVaugh, sp. nov. 

Arbor grandis, ramulis inflorescentibusque rufo-hirsutis, foliis 10-16 cm. 
longis, floribus 8-13, ultimis solitariis, infimis ternis, alabastris 5-6 mm. longis 
obovoideis vel subglobosis. 

A tree up to 20 meters high, named for its superficial resemblance to Krugia 
ferruginea, a native of the West Indies and northern South America. The few- 
flowered paired spikes, 5-6 cm. long, are conspicuously rufous-hirsute, and the 
buds completely obscured by the hairs; style 8 mm. long; stamens 125-150; 
petals 3, 4 mm. long in the bud. 

Amazonian Brazil: Basin of Rio Jurua, B. A. Krukoff 5041 
(NY; US). Peru, Loreto: Iquitos, edge of lake, elev. 120 meters, 
Oct. 11, 1929, L. Williams 3675 (F 618382, type). Univ. of Mich. 
Neg. 465. 

Calyptranthes maxima McVaugh, sp. nov. 

Arbor parva, puberulenta; ramulis paniculisque sparsiuscule pilis adpressis 
minutis, 0.1 mm. longis dibranchiatis, obsitis; foliis ellipticis maximis, 25-39 cm. 
longis, acuminatis; petiolo incrassato, 3 mm. diametro, 4 mm. longo; nervo medio 
supra elevato sulcatoque; paniculis paucifloris, floribus puberulentis furfurace- 
isque, maximis; alabastris 7-8 mm. longis; calyptra explanata, vix apiculata; 
staminibus circiter 200. 

A tree with leaves thin but extremely large for the genus, and the flowers 
large and coarse; leaves narrowly elliptic, acuminate, 3.5-4 times as long as wide, 
with 25-35 pairs of slender lateral veins; petiole 4 mm. long, its thickened portion 
extending also 5-6 mm. beneath the cordate-auriculate base of the blade; flowers 
up to 10 on each main branch of the panicle, arising from a stout flattened axis 
up to 3.5 mm. wide; stamens relatively short for the flower, 5 mm. long, fringing 
the summit of the hypanthium. 

Colombia, Amazonas: Trapecio amazonico, Loretoyacu River, 
elev. about 100 meters, Oct., 1946, R. E. Schultes & G. Black 8526 
(US 1996504, type). 

Calyptranthes plicata McVaugh, sp. nov. 

Frutex, foliis maximis obovatis, 30-38 cm. longis, sessilibus, marginibus ad 
basin decurrentibus plicatisque; paniculis tenuibus usque ad 7 cm. longis, pilis 
appressis runs, partim dibranchiatis obtectis; paniculae bracteis divaricatis per- 
sistentibus, 2-3 (-5) mm. longis; alabastris 3.5 mm. longis, obovoideis, acutis, 
sessilibus; staminibus circiter 60. 


A shrub about 4.5 meters high, with stem 5 cm. in diameter, nearly glabrous 
except the appressed-pubescent divaricately branched panicle which bears 50-100 
flowers; leaves 10-12 cm. wide, 2.5-3 times as long as wide, with 20-25 pairs of 
conspicuous lateral veins; a distinctive species because of the large sessile leaves 
with broadly decurrent and puckered basal margins. 

The inflorescence, with its persistently bracteate nodes, divari- 
cate branches and relatively few sessile flowers, suggests that of 
Calyptranthes longifolia Berg but is more slender and elongate; 
vegetatively, the two species are dissimilar. Known only from the 
type, which bears one inflorescence in young bud, and a second, 
detached inflorescence with very much younger, undeveloped buds. 

Brazil, Amazonas: Municipality Sao Paulo Olivenca, near Pal- 
mares, on terra firma, low land, high forest, Sept. 11-Oct. 26, 1936, 
B. A. Krukoff8432 (NY, type). 

Calyptranthes rufotomentosa McVaugh, sp. nov. 

Frutex, panicula rufo-tomentosa et foliis elongatis glabris; foliis oblanceolatis, 
15-21 cm. longis, nervo medio supra elevato, petiolo incrassato; panicula circiter 
50-flora, 6-7 cm. longa; hypanthio urceolato, ad florendi tempus 4 mm. longo. 

A shrub, glabrous except for the heavy dark red tomentum which completely 
covers the inflorescence; leaves narrow, and narrowly acuminate, 4-5 cm. wide, 
4 times as long as wide, the straight marginal vein and the rather widely spaced 
10-15 pairs of lateral veins conspicuous on the lower surface; flowers clustered 
toward the tips of the panicle branches; stamens about 75; style apparently 
undeveloped in the flowers examined; calyptra conic, apiculate, 1 mm. long. 

Brazil, Amazonas: Sao Paulo de Olivenca, caatinga de arvores 
baixas em areia branca, quasi pura, Oct. 15, 1942, A. Ducke 2240 
(NY, type). 

Calyptranthes sessilis McVaugh, sp. nov. 

Frutex, glaber, foliis sessilibus ovatis 9-12 cm. longis acuminatis; venis supra 
impressis; floribus in paniculis angustis perpaucis; alabastris fusiformibus 7 mm. 

Buds 2 mm. thick, the operculum 4-4.5 mm. long; style 8.5 mm. long; stamens 
60-75; fruit globose, 7 mm. in diameter. 

This species is similar to C. spruceana Berg, but in that species 
the buds are said to be globose and 5 mm. in diameter, the leaves 
are obtusely rounded at the tips, the veins are not impressed on the 
upper surface, and the fruits are more clustered toward the tips of 
the inflorescence. 

Peru, Loreto: Yurimaguas, L. Williams 4583 (F 623440, type), 
Killip & Smith 27603 (NY; US). Univ. of Mich. Neg. 446. 


Calyptranthes simulata McVaugh, sp. nov. 

Frutex vel arbor, pubescens; foliis ovatis lanceolatisve 9-15 (-25) cm. longis 
subglabris; venis utroque latere 20-25; paniculis binis e ramo compresso abortive 
1 cm. longo oriundis; paniculae ramis infimis usque ad 3-5 cm. longis; alabastris 
2-2.5 mm. longis obovatis, apice rotundatis vel brevi-apiculatis; hypanthio 

A shrub or tree to 9 meters high, belonging to the group of species with 
middle-sized and nearly glabrous leaves and many small flowers in much-branched 
panicles. The calyptra is explanate, 1.3-2 mm. wide; style 5-6 mm. long; stamens 

Known only from specimens in which the flowering panicles 
are borne in the terminal axils of the twigs, on old wood. The normal 
size and shape of the leaves on sterile twigs cannot be determined 
with certainty, nor can the presence or absence of wings on the 
branchlets be verified. It is possible that Calyptranthes simulata 
and C. tessmannii are conspecific, but on the basis of the flowering 
specimens at hand the two species are readily separable by the use 
of the key characters. 

Peru, Loreto: Iquitos, Aug. 2-8, 1929, Kittip & Smith 27352 
(NY; US), 26916 (F; NY; US); Iquitos, G. Tessmann 5372 (G; NY). 
Peru-Colombia boundary: forest near Rio Putumayo, Sept. 26-Oct. 
10, 1930, G. King 1614 (F; GH; MICH, type; NY; US). 

A collection from Tarapoto, San Martin, Peru, Wittiams 6539 
(F), may be the same species. It bears immature globose or oblate 
fruits about 1 cm. in diameter; the pubescence is exactly that of 
C. simulata and the leaves are similar in shape to those of that 
species. The specimen suggests C. tessmannii, however, in that the 
lateral veins hardly exceed 15 in number, and the panicle arises 
laterally from the base of a leafy shoot. 

Calyptranthes tessmannii Burret, in herb., ex McVaugh, 
sp. nov. 

Arbor vel frutex, pubescens; foliis ellipticis, 8-11.5 cm. longis, subglabris, 
venis utroque latere ca. 15; paniculis duabus solitariis, oppositis, e nodo infimo 
rami lateralis foliosi oriundis; paniculae ramis infimis 1.5 cm. longis; alabastris 
3-3.5 mm. longis fusiformibus, apice conspicue apiculatis; hypanthio strigoso. 

A species which appears distinctive because of the narrow panicles arising 
from the lowest node of a leafy shoot; the relatively narrow and fusiform buds 
with long-apiculate tips are also characteristic. The style is about 6 mm. long; 
stamens about 50. Flowering specimens are readily distinguished from Calypt- 
ranthes multiflora, C. ruiziana and C. simulata by the key characters. 

Peru, Loreto : Stromgebiet des Maranon von Iquitos . . . am 
Pongo de Manseriche, G. Tessmann 4832 in 1924 (G, type). F.M. 
Neg. 23395. 



1. Leaves sessile, cordate, 
la. Leaves ovate, 13-17 cm. wide, 1.7-2 times as long as wide, the veins impressed 

above; inflorescence decompound, stout, 20 cm. long or more; buds 7 mm. 

long or more; northeastern Peru (Loreto) . .M. obumbrans (Berg) McVaugh. 
la. Leaves lanceolate, 3.5-5 cm. wide, 2.5-4 times as long as wide, the lateral 

veins not impressed above; inflorescence few-flowered, 4-7 cm. long; buds 

4-6 mm. long; northern Bolivia (La Paz) M. connata McVaugh. 

1. Leaves definitely petiolate, cuneate or rounded at base or occasionally 

2. Inflorescence at anthesis (and usually in fruit) conspicuously bracteate, the 
bracts ovate, pointed, 6-12 mm. long; calyx-lobes 2.5-6 mm. long, lanceolate 
or ovate, much longer than wide; branchlets long-hirsute. 

3. Hairs of the branchlets reddish brown, up to 3.5 mm. long; calyx-lobes 4, 
the larger ones 6 mm. long M. huallagae McVaugh. 

3. Hairs of the branchlets yellowish brown, up to 2 mm. long; calyx-lobes 5, 
rarely 4, the larger ones 2.5-4 mm. long M. bracteata (Rich.) DC. 

2. Inflorescence with small inconspicuous bracts which are deciduous usually 
before the flowers open; calyx-lobes 3 mm. long or usually much less, rounded 
to subtruncate or triangular, mostly as wide as, or wider than, long; 
branchlets various. 

4. Summit of the ovary and interior of the prolonged and cuplike hypanthium 
glabrous; fruit usually globose, 5-6 mm. in diameter. [This includes 7 species 
which were assigned by Berg to Aulomyrcia, and one additional species, 
Myrcia bipennis (Berg) McVaugh, which is treated below.] 

4. Summit of the ovary and usually the interior of the hypanthium hairy; 
hypanthium variously or not at all prolonged beyond the ovary, the center 
of the flower thus flat or variously depressed; fruit various. 

5. Outer surface of the hypanthium glabrous, the entire plant glabrous or nearly 
so; fruiting hypanthium prolonged into a neck 2 mm. long at the apex of 
the fruit M. subglabra McVaugh. 

5. Outer surface of the hypanthium hairy, usually strongly so; fruiting hy- 
panthium, as far as known, not prolonged into a neck. 

6. Inflorescence slender, the axis often terete or, if compressed, mostly less than 
1.5 mm. wide (measured just below the lowest branches); small-flowered 
species, the disc 1-2 (rarely to 2.5) mm. wide; fruit ellipsoid or oblong, 5-9 
mm. long. [This includes M. poeppiyiana Berg, M. paivae Berg, M. splendens 
(Sw.) DC., M. sylvatica (Mey.) DC., and slender forms of M. fallax (Rich.) 

6. Inflorescence relatively stout, the axis and branches compressed or strongly 
flattened; axis at least 1.5 mm. wide and often 2.5-4 mm. wide (measured 
just below the lowest branches); flowers small or large; fruit various. 

7. Calyx in bud and in fruit truncate, the lobes one-third to one-half as long 
as wide, together forming a nearly straight-margined band or crown en- 
circling the summit of the hypanthium; hypanthium prolonged up to 2 mm. 
beyond the summit of the ovary. [Includes 2 species, both referred to 
Aulomyrcia by Berg: A. chilensis Berg, A. neesiana (DC.) Berg.] 

7. Calyx-margin not truncate, the lobes together not forming a recognizable 
unit but individually spreading, with evident sinuses between them; indivi- 
dual lobes broadly rounded or triangular, as long as wide, or a little wider 
than long; hypanthium various, usually not prolonged. 


8. Inner surface of the calyx-lobes glabrous or with a few scattered hairs, often 
with large glandular dots; small-flowered species, the disc 1.3-2 (rarely 2.5) 
mm. wide; style (as far as known) 5 mm. long or less. 

9. Upper leaf surface with 5-8 large convex glandular dots per square milli- 
meter; principal veins impressed M. perlusa DC. 

9. Upper leaf surface without apparent glands, or with minute impressed glands; 

veins not impressed, or the laterals or midvein somewhat so. 
10. Upper leaf surface thickly but minutely impressed-puncticulate; lateral veins 
(15-) 20-25 pairs; inflorescence often with the principal branches from near 
the base M. fascicularis Berg. 

10. Leaves not impressed-puncticulate; lateral veins 10-15 pairs; inflorescence 
usually not branching from near the base. 

11. Lateral veins impressed, 10-12 pairs, much stronger than the intermediate 
ones; midvein hairy, flat or somewhat convex above; inflorescence hirsute 
with yellowish hairs up to 1 mm. long; hypanthium elongate in bud, with 
about 10 narrow longitudinal ridges M. ambivalens McVaugh. 

11. Lateral veins not impressed, often slightly elevated in drying, about 12-15 
pairs, these often little stronger than the intermediate ones; midvein gla- 
brous or pubescent, sulcate or slightly impressed; inflorescence sparingly 
short-pubescent (rarely hirsutulous) ; hypanthium narrowly conic in bud, 
not ridged. [Includes M. fallax (Rich.) DC., and vigorous forms of M. 
splendens (Sw.) DC.] 

8. Inner surface of the calyx-lobes appressed-pubescent; small- or large-flowered 
species; style various. 

12. Lower leaf surface finely bullate; pubescence of inflorescence soft, ochraceous 
or rusty M. deflexa (Poir.) DC. 

12. Lower leaf surface smooth or verruculose, not bullate; pubescence various. 

13. Branches of the inflorescence, and the hypanthium, abundantly tomentose 
or hirsutulous with reddish hairs. 

14. Inflorescence rufous-tomentose; leaves 10-15 cm. long, broadly and bluntly 
acuminate, or obtusely pointed, lustrous and nearly glabrous beneath, with 
prominent large glands M. atrorufa McVaugh. 

14. Inflorescence pale rufous-hirsutulous with hairs of different lengths inter- 
mixed; leaves 12-26 cm. long, abruptly acuminate, dull and pale beneath, 
eglandular but with numerous appressed hairs. .M. egensis (Berg) McVaugh. 

13. Branches of the inflorescence, and the hypanthium, with white or yellow, 
appressed or spreading, mostly straight hairs; hairs sometimes confined to 
the hypanthium and then occasionally reddish or tawny. 

15. Lateral veins, and often the marginal vein also, impressed on the upper 
leaf surface [see also M. madida, a large-leaved species in which the lateral 
veins may dry as fine lines slightly below the surface level]. 

16. Leaves hirsute beneath, at least on the veins, with soft yellow hairs up to 
1 mm. long; inflorescence similarly hirsute with spreading or ascending hairs; 
lateral veins about 10 (rarely -15) pairs; disc 3-3.5 mm. wide. 

M. mollis (HBK.) DC. 

16. Leaves sparingly appressed-pubescent beneath, or with a few long hairs on 
the veins; inflorescence sparingly hispidulous to appressed-pubescent with 
short hairs; lateral veins mostly more than 15 pairs. 

17. Leaves rounded, obtuse, or emarginate at apex, middle-sized (mostly less 
than 15 cm. long), relatively broad (mostly not more than twice as long as 

17a. Calyx-lobes about 1 mm. long; leaves rounded or emarginate at apex, 1.3-1.5 
times as long as wide; inflorescence and branchlets sparingly hairy. 

M . fasciata McVaugh. 


17a. Calyx-lobes 2-2.5 mm. long; leaves rounded to obtuse or obscurely apiculate, 
about twice as long as wide; inflorescence and branchlets velutinous. 

M. dispar McVaugh. 

17. Leaves markedly acuminate or acute, mostly larger (16-27 cm. long), rela- 
tively narrow (mostly 2-4 times as long as wide). 

18. Leaves bullate above, between the 15-20 pairs of ladder-like veins; petiole 
1.5-2 mm. long; lateral veins prominent beneath near the mid vein only; 
small-flowered species, the disc 1.5-2 mm. wide M. fenestrata DC. 

18. Leaves flat or nearly so in drying; petioles 3-8 mm. long; lateral veins promi- 
nent their whole length; large-flowered species, the disc 3-4 mm. wide. 

19. Leaves appressed-pubescent beneath; midvein sulcate above; flowers aggre- 
gated on very short angular branchlets; hypanthium strongly angled in 
drying, whitened by thick felted hairs M. pentagona McVaugh. 

19. Leaves nearly glabrous and markedly verruculose beneath; midvein forming 
a narrow ridge above; inflorescence loosely branched, the flowers mostly in 
threes on slender branchlets; hypanthium conic, strigose. 

M. crassimarginata McVaugh. 

15. Lateral and marginal veins not impressed on the upper surface, usually 
slightly raised in drying. 

20. Leaves hirsute beneath, at least on the veins, with soft yellow hairs up to 
1 mm. long; inflorescence similarly hirsute with spreading or ascending hairs; 
disc (2.5-) 3-3.5 mm. wide; lateral veins about 10 (rarely -15). 

M. mollis (HBK.) DC. 

20. Leaves appressed-pubescent or nearly glabrous beneath; inflorescence ap- 
pressed-pubescent or short-hirsutulous with hairs less than 0.5 mm. long; 
disc and veins various. 

21. Flowers aggregated near the tips of the panicle branches, those of the terminal 
clusters sessile or on very short pedicels up to 1 mm. long and nearly as thick; 
disc 3 mm. wide or less; lateral veins 15-20 pairs. 

22. Hypanthium whitened with very short pale felted hairs; disc 2.5-3 mm. 
wide; leaves 18-23 cm. long; buds 4-4.5 mm. long. . . .M. madida McVaugh. 

22. Hypanthium covered thinly by appressed short silky hairs; disc 2-2.5 mm. 
wide; leaves 7-14 cm. long; buds 3 mm. long M. concava McVaugh. 

21. Lateral flowers of the terminal clusters on slender pedicels 2-6 mm. long; 
disc 3-4 mm. wide; leaves 16 cm. long or less, with 10-15 pairs of lateral veins. 

23. Branches of the inflorescence short-hirsutulous; marginal vein about as 
strong as the laterals and strongly arched between them. 

M. albobrunnea McVaugh. 

23. Branches of the inflorescence appressed-strigose or nearly glabrous; lateral 
veins diminishing distally and usually not forming a distinct marginal vein. 

M. dichasialis McVaugh. 

Myrcia albobrunnea McVaugh, sp. nov. 

Arbor, ochraceo-hirsutula, foliis acuminatis 9-16 cm. longis; venis supra vix 
manifestis; nervo marginali prominente; paniculis multifloris, rhachi valida 1.5- 
2.5 mm. lata, floribus ultimis tenuiter pedicellatis; calycis lobis intus extusque 
pubescentibus; germine hirsute, disco 4 mm. lato. 

A tree 3-8 meters high, the ovate leaves with 10-15 pairs of lateral veins; 
flowers large, the style 7.5-8 mm. long, the stamens more than 200. Flowers 
(according to Klug) white and brown. The center of the flower is much depressed 
(i.e., the hypanthium is prolonged about 1 mm. beyond the summit of the ovary), 
as in the supposed genus Aulomyrcia. 


This species strongly resembles M. dichasialis, which is a native 
also of the vicinity of Iquitos, and it is with some hesitancy that 
I describe them both as new. Myrcia albobrunnea differs from 
M. dichasialis in having the branchlets markedly hirsutulous in- 
stead of glabrous or sparingly appressed-strigose. The flowers in 
M. albobrunnea are slightly larger, the leaves are less markedly 
reticulate on the upper surface, the midvein is hairy above, and the 
marginal vein is well defined although strongly arched between the 
laterals; the lateral flowers of the terminal triads are somewhat 
more conspicuously pedicellate in M. dichasialis. 

Peru, Loreto: Iquitos, elev. 100 meters, woods, Killip & Smith 
27006, Aug. 3-11, 1929 (F; US) ; Mishuyacu, elev. 100 meters, forest, 
Feb.-Mar., 1930, G. King 1030 (F; US 1456120, type). 

Myrcia ambivalens McVaugh, sp. nov. 

Frutex, ramulis paniculisque pilis flavidis, flexuosis, usque ad 1 mm. longis 
obsitis; foliis 5-12 cm. longis, acuminatis; paniculis multifloris validis; nervo 
medio supra piano vel vix elevato; venis utroque latere 10-12, supra impressis; 
calycis lobis intus glabris; disco piloso, ut videtur 2 mm. lato; hypanthio circiter 

Leaves ovate, elliptic or obovate, nearly concolorous and eglandular, slightly 
lustrous above. The type is in bud; the flowers are small: buds 4 mm. long, the 
calyx-lobes 1-1.5 mm. long, the stamens about 200. 

A distinctive species because of the indument, the comparatively 
strong and few veins of the leaves, and the elongate and sulcate 

Peru, Loreto: Yurimaguas, in forest, Nov. 7, 1929, L. Williams 
4706 (F), 4737 (F 624825, type). Univ. of Mich. Neg. 473. 

Myrcia atrorufa McVaugh, sp. nov. 

Frutex, rufo-pubescens vel -tomentosus; foliis obtuse acuminatis vel obtusis, 
subtus lucidis, conspicue glandulosis; paniculis validis, rhachi usque ad 2.5 mm. 
lata; calycis lobis intus pubescentibus; disco 2.5-3 mm. lato, piloso, concavo. 

A shrub 2 meters high with ovate leaves 10-15 cm. long, the foliage distinctive 
because of the usually bluntly pointed tips of the blades, the coarse reticulum 
formed by the small veins on the upper surface, and the prominent large glands 
beneath. The crisped rufous hairs of the branchlets and inflorescence are unlike 
those of other Peruvian species. Style probably 4 mm. long; stamens 125-150; 
fruit subglobose, about 1 cm. in diameter. 

Peru, Huanuco: Between Huanuco and Pampayacu, R. Kanehira 
17, Jan. 30, 1927 (A; F; GH). Puno: Prov. Carabaya, trail from 
Santo Domingo to Chabuca mine, elev. 1,900 meters, moist open 


places with dense vegetation, much fog and rain, R. D. Metcalf 
30671, May 30-June 1, 1942 (US 1876047, type). 

Myrcia bipennis (Berg) McVaugh, comb. nov. Myrciaria 
bipennis Berg, Linnaea 31: 259. ?1862. 

A most distinctive species, but quite out of place in the genus 
to which Berg assigned it. The seeds in the recent collections made 
in Amazonian Brazil by Holt & Blake (no. 556) and in Venezuela by 
Williams (no. 14763) are definitely myrcioid. No flowering material 
was available to Berg, and apparently none has been collected since 
his time; the immature buds on the type specimens (Spruce, no. 
3770) suggest the genus Marlierea, and the morphology of the 
inflorescence suggests Marlierea more than Myrcia. The hypan- 
thium on the mature fruit, however, is straight-sided within and 
with a definite rim (as in the Aulomyrcia group of Myrcia), and the 
calyx-lobes are well formed, not irregularly split at base as in 

Myrcia concava McVaugh, sp. nov. 

Arbor, paniculis multifloris, strigosis, 7-11 cm. longis, rhachi valida, usque 
ad 2 mm. lata; foliis acuminatis, 7-14 cm. longis; venis utrinque prominulis, non 
impressis; floribus subsessilibus, germine piloso, disco 2-2.5 mm. lato, hypanthio 
extus sericeo-strigoso, calycis lobis intus pubescentibus. 

A tree 6-8 meters high with elliptic or ovate leaves about 2.5 times as long 
as wide; buds 3 mm. long; hypanthium-base not attenuate, but abruptly con- 
tracted into the pedicel; style 4.5 mm. long; stamens about 150. 

A species which is separated by no very obvious characters from 
those of the Myrcia fallax-splendens complex, but which differs from 
these in its short, stout inflorescence, its conspicuous calyx-lobes, 
which are pubescent within and concave and spreading as the petals 
fall, and its sparingly strigose and broad-based hypanthium. 

Peru, Loreto: Mishuyacu, near Iquitos, forest, elev. 100 meters, 
G. King 454 (F; NY; US), 800 (F; NY; US 1455791, type). 

Myrcia connata McVaugh, sp. nov. 

Frutex 5-6 m. altus, strigosus; foliis sessilibus lanceolatis, 9-16 cm. longis, 
cordatis; nervo medio supra impresso; venis utroque latere 12-15, inconspicuis; 
paniculis paucifloris validis; calycis lobis intus pilosis; disco 3 mm. lato, piloso, 

A shrub 5-6 meters high, the branchlets, leaf-buds and inflorescence (es- 
pecially the calyx and hypanthium) strigose with straight yellowish hairs up to 
0.5 mm. long; at least the young leaves sparingly strigose beneath; leaves sessile, 
lanceolate, 3.5-5 cm. wide, 9-16 cm. long, about 2.5-4 times as long as wide, acute 


or obscurely acuminate, blunt at very tip, cordate at base, the auricles of opposite 
pairs overlapping; midvein impressed above, prominent beneath; lateral veins 
12-15 pairs, slender, inconspicuous on both sides, more prominent beneath; 
marginal vein about equaling the laterals and arched between them, 3-5 mm. 
from the margin; small veins markedly reticulate in prevailingly right-angled 
patterns; blades somewhat lustrous above, paler and dull beneath, the glandular 
dots scarcely apparent; inflorescence a panicle 4-7 cm. long, about 12-flowered, 
once- or twice-compound, the peduncle 2.5-4.5 cm. long, 2 mm. wide below the 
first node; lower branches of the panicle up to 1.2 cm. long; flowers sessile; buds 
4-6 mm. long, the hypanthium about 2 mm. long, heavily appressed-hairy; 
calyx-lobes 5, broadly rounded, appressed-hairy on both sides, 1.3-1.5 mm. long, 
2.5 mm. wide; disc about 3 mm. wide, sunken, the hypanthium prolonged 1-1.5 
mm. beyond the summit of the ovary; style probably about 10 mm. long; stamens 
150-200, 7-8 mm. long, the anthers 0.6 mm. long; petals hairy outside, about 
7-8 mm. long and 5 mm. wide; fruit fleshy, about 1 cm. in diameter, a little longer 
than wide. 

Bolivia, La Paz: Prov. S. Yungas, basin of Rio Bopi, San Barto- 
lom< (near Calisaya), elev. 750-900 meters, July 1-22, 1939, B. A. 
Krukoff 10382 (NY, type; US; Y). 

Myrcia crassimarginata McVaugh, sp. nov. 

Arbor, ramulis paniculisque pilis brevibus ochraceis puberulentis, ramulis 
novellis velutinis; foliis acuminatis, subtus minute siliceo-verruculosis, venis 
impressis; nervo marginali prominente, nervo medio supra elevato planiusculo; 
paniculis multifloris validis, rhachi usque ad 3 mm. lata; floribus ultimis tenuiter 
pedicellatis; germine piloso; disco 3-4 mm. lato; calycis lobis intus strigosis. 

A tree 5 meters high, the leaves elliptic, 16-20 cm. long, 2-2.5 times as long 
as wide; the foliage is characteristic because of the prominent and strongly im- 
pressed veins, and the ridge formed by the midvein on the upper leaf-surface. 
Flowers large, the buds 5 mm. long; style 6-8 mm. long; stamens about 300. 

A collection made by Tessmann (no. 4319), at the mouth of 
Rio Santiago, on high land, is apparently the same species, but in 
the specimen seen (at NY) the leaves are appressed-pubescent and 
smooth beneath (not verruculose) . 

Peru, Loreto: Fortaleza, near Yurimaguas, forest, elev. about 
140 meters, Dec., 1932, G. King 2821 (A, type; F; G; GH; NY; US). 

Myrcia dichasialis McVaugh, sp. nov. 

Arbor, foliis ovatis brevi-acuminatis 8-15 cm. longis, venis supra non impressis, 
longiori a margine distantia arcuatim unitis; paniculis multifloris validis, rhachi 
2-2.5 mm. lata, ramulis appresse-strigosis vel subglabris, floribus ultimis tenuiter 
pedicellatis; germine hirsute; disco 3-4 mm. lato; calycis lobis intus pubescentibus. 

Perhaps a shrub, or a tree up to 10 meters high, with ovate leaves 2-3 times 
as long as wide, the marginal vein usually not distinct from the laterals which 
diminish distally as they arch inward and fuse with the next succeeding ones. 
Flowers large; style about 7 mm. long; stamens about 200. 


The distinctions between this species and M. albobrunnea are 
mentioned above, in the discussion following the description of 
that species. 

Peru, Loreto: Iquitos, Mexia 6508 (F; G; US). Gamitanacocha, 
Rio Mazan, J. M. Schunke 134 (A; F 997268, type; US), 188 (A; 
F; US). Stromgebiet des Ucayali von 10 S. bis zur Miindung, 
G. Tessmann 3411 (G). Along Rio Itaya, L. Williams 91 (F). 

Myrcia dispar McVaugh, sp. nov. 

Arbor velutina, paniculis ramulisque dense pilis aureo-flavidis usque ad 0.8 
mm. longis obtectis; foliis obtusis vel breve apiculatis, 10-15 cm. longis, subtus 
minute siliceo-verruculosis strigosisque; venis impressis; nervo marginali supra 
non impresso; paniculis circiter ut videtur 20-floris, validis, rhachi usque ad 2.5 
mm. lata; floribus sessilibus; germine piloso; disco 3 mm. lato; calycis lobis intus 
strigosis, 2-2.5 mm. longis. 

A tree said by Krukoff to be 45 feet high; leaves coriaceous, elliptic to ovate 
or obovate, very inconspicuously if at all acuminate, about twice as long as wide; 
panicles stout and few-flowered, notably shorter than the leaves, even in fruit 
retaining considerable amounts of the indument which covers the branchlets; 
fruit globose, 1-1.5 cm. in diameter, crowned by the erect calyx; flowers not seen 
but apparently large for the genus. 

Differs from other known species in the short, stout inflorescence, 
relatively large and few flowers, coarsely veined blunt leaves, and 
copious distinctive indument. It is associated in the key with 
Myrcia fasciata, but it bears little resemblance to that species 
except in the key characters. Its actual relationships are not 
apparent to me. 

Brazil, Acre: Basin of Rio Jurua, upper Rio Jurupary, Lat. 
8-9 S., Long, about 70 W., on terra firma, July 15, 1933, B. A. 
Krukoff 5232 (NY, type). 

Myrcia egensis (Berg) McVaugh, comb. nov. Aulomyrcia 
egensis Berg, in Mart. Fl. Bras. 14, pt. 1: 99. 1857. A. macrophylla 
Berg, in Mart. I.e. 

The type of A. egensis, Poeppig's no. 2551, from Ega, Brazil, 
bears immature fruit. Except for somewhat shorter petioles it 
seems not to differ significantly from a flowering specimen collected 
by Klug (no. 3569) in Peru. I have not made a direct comparison 
between the type of A. macrophylla and more recently collected 
materials, but from notes made upon the type in Munich in 1954, 
and from a photograph of it (F.M. Neg. 19818), I believe it to be 
conspecific with the other two collections cited above. 


Myrcia fascia ta McVaugh, sp. nov. 

Arbor strigosa, foliis rotundo-ovatis, coriaceis, venis impressis; paniculae 
ramulis compressis, pedunculo superne usque ad 3-4 mm. lato; germine piloso; 
disco circiter 2.5 mm. lato; hypanthio pilis plurimis atro-fuscis obsito; calycis 
lobis intus pubescentibus. 

A tree to 12 meters high, with lustrous coriaceous broad leaves which are 
rounded or emarginate at apex; flowers about 25, in small groups near the tips 
of the much-flattened panicle branches; mature flowers not seen, but the buds 

2 mm. long; fruit subglobose, about 1 cm. in diameter. 

Known only from the following collections, one of which bears 
very immature buds, and the other nearly mature green fruit. 

Ecuador, Azuay: Forested slopes between Cruz Pamba and Loma 
de Canela, in region of Rio Sadracay, tributary of Rio Mehuir, 
north of Molleturo, elev. 2,315-2,500 meters, June 12, 1943, J. A. 
Steyermark 52969 (F 1391179, type), 52961 (F). 

Myrcia huallagae McVaugh, sp. nov. M. lanceolate, 7 grandi- 
folia Berg, in Mart. Fl. Bras. 14, pt. 1: 155. 1857. 

Frutex debilis, hirsutus, pilis rufis erectis usque ad 3.5 mm. longis obsitus; 
foliis elliptico-lanceolatis, basi subauriculatis; floribus paucis, bracteis bracteo- 
lisque persistentibus, ovatis, usque ad 12 mm. longis, suffultis; calycis lobis 4, 
exterioribus majoribus; disco concavo, 3.5 mm. lato; M. bracteatae maxime affinis. 

A shrub or woody vine up to 3-4 meters high, with leaves 8-12 cm. long, often 

3 times as long as wide; the raceme-like inflorescences are up to 4 cm. long, much 
shorter than the leaves; style 6 mm. long; fruit ellipsoid, 13 mm. long. 

Closely akin to Myrcia bracteata and scarcely to be distinguished 
from that species except by the characters given in the key and by 
the somewhat larger size of most vegetative and reproductive parts. 
It is here recognized as a species because it is separated from M. 
bracteata by a series of tangible though perhaps minor features, and 
because it occupies a restricted geographical range. 

Peru, Loreto: Yurimaguas, Kittip & Smith 27976 (US), 28188 
(US), L. Williams 3837 (F), 3888 (F), 4715 (F). In sylvis ad Hua- 
llagam, Feb., 1831, Poeppig 2267 (W, type, and type of M. lanceo- 
lata 7 grandifolia) . 

Myrcia tnadida McVaugh, sp. nov. 

Arbor, minute puberula, foliis acuminatis, 9-23 cm. longis; venis utrinque 
prominulis, non impressis; paniculis multifloris, validis, rhachi compressa, usque 
ad 3.5 mm. lata; floribus subsessilibus, germine piloso, disco 2.5-3 mm. lato, 
hypanthio extus basin versus pilis brevissimis incano-sympilematis obtecto, 
calycis lobis intus pubescentibus. 

A tree to 12 meters high, nearly glabrous but with very small appressed pale 
hairs which are numerous and densely aggregated on the leaf-buds and the base 


of the hypanthium, whitening the surface and in drying somewhat felted together 
as if from moistening (madida, drenched, sodden). Leaves elliptic to ovate or 
obovate, about 3 times as long as wide; style 6 mm. long; stamens about 125; 
fruit black, ellipsoid, 10-13 mm. long. 

Known only from the following collections: 

Peru, Loreto: Rio Napo near Mazan, elev. 110 meters, over- 
hanging river, Mexia 6448, Jan. 27, 1932 (F 718476, type; G; US). 
Manfinfa, upper Rio Nanay, L. Williams 1088 (F). 

Myrcia obumbrans (Berg) McVaugh, comb. nov. Rubachia 
obumbrans Berg, in Mart. Fl. Bras. 14, pt. 1: 28. 1857. Marlierea 
obumbrans (Berg) Niedz. in Natiirl. Pflanzenfam. Ill, pt. 7: 76. 1893. 

Berg assigned this species to Rubachia because of his observation 
that after an thesis the calyx split longitudinally below the sinuses. 
I have not been able to confirm this, but the plant bears marked 
resemblances to other Peruvian species of Myrcia and seems to have 
little in common, either morphologically or geographically, with the 
6 south-Brazilian species of Rubachia. The genus Rubachia is in- 
deed of doubtful validity; Bentham (Benth. & Hook. f. Gen. PI. 1: 
717. 1865) relegated it to the synonymy of Marlierea; Niedenzu, as 
noted above, made the formal transfer of Rubachia obumbrans to 
Marlierea. On the basis of the material which I have seen (Killip 
& Smith 28180; Williams 4581), I judge that the calyx in this species 
is that of Myrcia, not of Marlierea. These specimens, however, are 
in bud only, and additional collections in flower or young fruit are 
needed to confirm these observations. 

Myrcia pentagona McVaugh, sp. nov. 

Arbor, ramulis, paniculis et praesertim hypanthio, pilis mollibus sympilematis, 
albicatis; foliis acuminatis, 20-30 cm. longis, venis impressis; paniculis multifloris, 
validis, rhachi usque ad 4 mm. lata; floribus sessilibus, e ramulis subtetragonis; 
germine piloso, 3-4 mm. lato; hypanthio 5-gono; calycis lobis intus pubescentibus. 

A tree 5 meters high, with elliptic leaves 3-4 times as long as wide and 18-25 
pairs of strong lateral veins and a strongly impressed marginal vein. The peculiar 
indument and the strongly angled hypanthium are characteristic. Flowers large: 
style 6-7 mm. long; stamens about 150. 

Peru, Loreto: Florida, Rio Putumayo, mouth of Rio Zubineta, 
elev. 180 meters, forest, May-July, 1931, G. King 2152 (F 668840, 
type; G; GH; NY; US). 

Myrcia splendens (Sw.) DC., var. chrysocoma McVaugh, 
var. nov. 


Arbor, M. splendenti maxime affinis, sed distincta: Indumento densiore, 
longiore, pilis appressis vel ascendentibus, lucidis, sericeis, fulvis vel ochraceis, 
usque ad 1 mm. longis, calycis lobos superantibus. 

This is the silky-pubescent extreme of the forms of Myrcia 
splendens; a somewhat less pubescent form is M. sericea Berg, and 
another is the silky-strigose M. saxicola Berg. The present variety 
is known only from the type collection: 

Peru, San Martin: Pongo de Cainarachi, Rio Cainarachi (a tribu- 
tary of the Rio Huallaga), elev. 230 meters, in forest, Sept.-Oct., 
1932, G. King 2622 (A; F; G; GH; NY; US 1457061, type). 

Myrcia subglabra McVaugh, sp. nov. 

Arbor, subglabra (gemmis petalisque exceptis), foliis 8-20 cm. longis, apice 
graduatim acuminatis, basi marginibus decurrentibus; nervo medio supra im- 
presso, venis 15-18 supra planiusculis vel elevatis; laminis utrinque lucidis, grosse 
pellucido-punctatis; paniculis multifloris validis, rhachi 2 mm. lata; germine 
pallide piloso; hypanthio glabro, supra germen valde (2 mm.) producto, urceolato; 
calycis lobis utrinque glabris. 

A tree 6 meters high, with strongly flattened branchlets and leaves elliptic- 
lanceolate or narrowly ovate, 2.5-4 times as long as wide. The plant is readily 
identified by the prominently glandular and reticulately veined leaves and by the 
narrowly prolonged neck of the hypanthium, which exceeds in length those of all 
other species of Myrcia known to me. Flowers relatively large: disc 2.5-3 mm. 
wide; style 6-7 mm. long; stamens about 200; fruit ellipsoid-oblong, about 1 cm. 

Thus far known only from the Mapiri region of northern Bolivia, 
where it has been collected several times in the subtropical forested 
areas, at elevations of 570-850 meters. 

Bolivia, La Paz: San Carlos, Nov. 18, 1926, 0. Buchtien 956 
(F 928708, type; GH; NY; US), 942 (F; GH; NY; US), 943 (F; 
MICH; US), 944 (US), 957 (US); Charopampa, Buchtien 74 (F; G), 
1909 (G); San Antonio, Buchtien 1909 (US); Copacabana, Krukoff 
11099 (F). 

4. EUGENIA L. 1 

1. Leaves ternate, or opposite at some nodes, 2 cm. long or less; flowers solitary, 
axillary; northern Peru (Cajamarca) to Colombia E. triquetra Berg. 

1. Leaves opposite, or occasionally sub-opposite. 

2. (See second "2," p. 200.) Principal axis of the inflorescence racemosely 
branched, the branches or flowers in decussate pairs or groups; terminal 
flower of the axis usually wanting; axis elongate to extremely shortened, so 

1 Three species, extra-Peruvian in range, are described below but are not 
included in the key. These are Eugenia percrenata (Mato Grosso, Brazil), 
E. valvata (central Ecuador) and E. variareolata (eastern Colombia). 


that the flowers may appear to be racemose, or in axillary fascicles, umbels 
or glomerules; flowers, if occasionally solitary, arising from the base (the 
lowest radial node) of an abortive axillary bud or from the basal bracteate 
nodes of new branchlets which are leafy above; bracteoles at base of flower 
usually broad and persistent, often connate. 

Racemes irregularly compound, the primary branches cymosely 3 (or 
rarely 7)- flowered, or 1-flowered; hypanthium on a slender stipe (pseudo- 
stalk) often longer than itself, the linear deciduous bracteoles thus well below 
the flower, and the central flower of the cymose clusters apparently pedi- 
cellate E. stipitata McVaugh. 

Racemes simple, the primary branches (pedicels) 1-flowered; rarely the 
racemes racemosely compound, or with 2-3 flowers arising together from 
a node of the axis, or with 1-2 additional, subordinate racemes arising from 
the lowest (tangential) nodes of the primary one; hypanthium not slender- 
stipitate, the bracteoles close to and often surrounding the base of the flower. 
(See second "4," p. 196.) Axis of the raceme relatively slender and elongate, 
3-5 cm. long or more; if shorter, then the internodes much longer than the 
diameter of the axis, and some or all of them 5 mm. long or more. 
Axis of the raceme 1.5 cm. long or less, the ascending slender pedicels often 
as long as, or longer than, the axis; inflorescence appressed-strigose or -silky. 
[Includes E. biflora (L.) DC., E. inundata DC., and E. macrocalyx (Rusby) 

Axis of the raceme 2 cm. long or more or, if shorter, the pedicels 4-7 mm. 
long, widely spreading and much shorter than the axis; pubescence and 
foliage various, the leaves neither mucronate nor small and obovate-cuneate. 
Hypanthium glabrous without, the plants nearly glabrous, usually finely 
pubescent in the inflorescence; small-flowered species with slender, loosely 
flowered racemes. 

Lateral veins of the leaves 12-15 pairs, wide-spreading and nearly straight, 
the marginal vein nearly straight between the laterals and amply differenti- 
ated from them, 2-4 mm. from the margin; inflorescence closely appressed- 
pubescent, the calyx-lobes and bracteoles fimbriate-ciliate. 

E. calva McVaugh. 

Lateral veins 10 pairs or fewer, curved and ascending, at least the basal ones 
diminishing distally and not forming a definite marginal nerve; inflorescence 
glabrous, or finely hispidulous, or with minute appressed dibranchiate hairs; 

calyx-lobes and bracteoles sparsely short-ciliate E. florida DC. 

Hypanthium, at least the base, thickly strigose or pubescent without; plants 
variously pubescent, the inflorescence usually markedly so; flowers and 
racemes various. 

Leaves rounded or obtuse at tips, broadly elliptic-ovate, 7 cm. long or less; 
plants densely ferruginous-tomentose; flowers up to 4 pairs in a stout raceme 
2 cm. long, or some flowers solitary near the base of new leafy shoots. 

E. racemiflora Berg. 

Leaves usually markedly acuminate, relatively narrow and more than 7 cm. 
long; plants various, not ferruginous-tomentose (in E. atroracemosa rufous- 
velutinous); flowers all racemose. 

Bracts, bracteoles and calyx-lobes thickly beset with dark raised rounded 
glands, but inconspicuously if at all pubescent. 

Pubescence of appressed reddish-brown, partly dibranchiate hairs, intermixed 
with shorter erect hairs; connate bracteoles forming an involucre-like cupule 
2.5 mm. across; larger calyx-lobes 2.5 mm. long; style 5-6 mm. long; stamens 
about 100; coastal Ecuador (and Peru?) E. pustulescens McVaugh. 


10. Pubescence of appressed grayish-white hairs; cupule 4 mm. across; calyx- 
lobes 4 mm. long; style 8 mm. long; stamens 250-300; Amazonian Peru. 

E. polyadena Berg. 

9. Bracts, bracteoles and calyx-lobes not at all glandular or inconspicuously 
so, but variously and often conspicuously pubescent. 

11. Midvein elevated above in a narrow median line; inflorescence slender, the 
axis mostly not more than 1 mm. thick, closely appressed-pubescent with 
reddish brown mostly dibranchiate hairs E. dibranchiata McVaugh. 

11. Midvein impressed; inflorescence various. 

12. Axis of the raceme slender, 0.5-1 mm. thick (measured just below the lowest 
node); leaves mostly less than 10 cm. long; flowers small, the disc 2 mm. 
wide or less. 

13. Leaves tomentose beneath, sub-vernicose and rough above, with many 
glandular dots E. curvipilosa McVaugh. 

13. Leaves glabrous or essentially so at maturity, the upper surface smooth, 
rather dull, eglandular E. limbosa Berg. 

12. Axis of the raceme stouter, 1-2 mm. thick (measured just below the lowest 
node); leaves mostly more than 10 cm. long; flowers larger, the disc 2.5 mm. 
wide or more. 

14. Lateral and marginal veins impressed above, prominent beneath. 

15. Lateral veins 8-12 pairs; leaves caudate-acuminate; disc 6 mm. wide; veins 
of the lower leaf surface appressed-pubescent with tawny hairs. 

E. longicuspis McVaugh. 

15. Lateral veins about 15-20 pairs; leaves moderately acuminate; disc 3-3.5 
(-6) mm. wide; veins of the lower leaf surface velutinous with erect rufous 
hairs E. atroracemosa McVaugh. 

14. Lateral veins (about 15-20 pairs) and marginal vein elevated slightly but 
inconspicuous on both surfaces, the blades flat; disc 2.5-3 mm. wide; lower 
leaf surface, including veins, with very short, closely appressed colorless 

hairs E. riparia DC. 

4. (See first "4," p. 195.) Axis of the raceme much abbreviated, 2 cm. long or 
usually much less; if more than 1 cm. long the nodes approximate and the 
internodes 3 mm. long or less, and hardly longer than the thickness of the 
stout angled rachis; flowers sometimes solitary at the lowest nodes of new 
leafy branchlets, and in axillary clusters or short racemes on the same plant. 

16. Leaves markedly bullate, large and broad (5-10 cm. wide), elliptic to obovate; 
racemes mostly 10-18 mm. long, the axis quadrangular, with up to 8-13 
pairs of slender-pedicellate flowers E. tetrasticha Berg. 

16. Leaves flat, or the principal veins impressed above; blades variously shaped. 

17. Outer corky layers of the petiole irregularly loosening and flaking off, the 
whole much roughened and appearing 3-5 mm. thick; leaves large, 20-35 
cm. long, 3-5 times as long as wide; flowers clustered on old wood, the axis 
4 mm. long or less. 

18. Leaves mostly oblanceolate, cordate-auriculate and subsessile (petiole 3-5 
mm. long and about as thick); lateral veins about 15; midvein impressed 
above; calyx-lobes 4-5 mm. long E. multirimosa McVaugh. 

18. Leaves elliptic-oblong, rounded at base, the petiole 10-18 mm. long and 3-4 
mm. thick; lateral veins 20-30; midvein convex above; calyx-lobes 6-8 
mm. long E. tumulescens McVaugh. 

17. Petiole smooth or wrinkled, not thick and exfoliating, rarely more than 
2 mm. thick; leaves and flowers various. 

19. Large-leaved, large-flowered, coarse species (leaves 23-35 cm. long or more; 
petioles 2-4 mm. thick; buds 1-2 cm. long; anthers 1-1.8 mm. long). 


Lateral veins 25-30 pairs; marginal vein about as strong as the laterals; 
leaves about 4 times as long as wide; buds probably 2 cm. long, the sub- 
orbicular calyx-lobes up to 15 mm. long; bracteoles elliptic, probably de- 
ciduous E. scalariformis McVaugh. 

Lateral veins 9-12 pairs, prominent and ascending but diminishing distally 
and scarcely forming a marginal vein; leaves 2.3-3 times as long as wide; 
buds about 1 cm. long, the triangular or oblong calyx-lobes 4-6 mm. long; 
bracteoles obovate, enveloping the bud, their narrow connate bases forming 

a collar about the pedicel E. myrobalana DC. 

Leaves usually smaller, or with a definite marginal vein; flowers various, the 
buds mostly less than 1 cm. long; bracteoles usually small and persistent. 
Calyx-lobes foliaceous, elongate, erect, 7-9 mm. long at anthesis, separated 
by broad rounded sinuses; leaves 9-18 cm. long, with about 10 pairs of veins; 

anthers 0.3 mm. long E. macrocalyx (Rusby) McVaugh. 

Calyx-lobes not or scarcely foliaceous, mostly broad, concave and imbricate; 
if more than 6 mm. long the margins variously connate or overlapping and 
the anthers 1-1.5 mm. long; leaves various. 

Leaves with a glabrous cartilaginous margin (visible from above), this 
formed by a heavy convex vein about twice as thick as the lateral veins; 
blades elliptic-oblong, 6-8.5 cm. wide, 12-14 cm. long; plants velutinous 
with coarse reddish-brown hairs; flowers in sessile glomerules. 

E. percincta McVaugh. 

Leaves without a heavy vein at the margin; plants not coarsely rufous- 
velutinous; flowers various. 

Calyx-lobes large (the larger ones 5-8 mm. long or more), often as broad as 
long, much imbricated but distinct from the first, glabrous within; flowers 
probably always in sessile clusters on old wood. 

Inflorescence minutely appressed-pubescent with glistening brown hairs; 
leaves papillose-roughened above; style 9-10 mm. long. 

E. tenuimarginata McVaugh. 

Plants glabrous except the ciliate margins of the perianth lobes and bracte- 
oles; leaves smooth above; style 15 mm. long or more. 
Leaves 8-14 cm. wide and about 1.5 times as long; disc 4-5 mm. wide; anthers 

2-2.3 mm. long E. chartacea McVaugh. 

Leaves 4-7.5 cm. wide and 3-3.5 times as long; disc 6-7 mm. wide; anthers 

1-1.3 mm. long E. illepida McVaugh. 

Calyx-lobes smaller (the larger ones 6 mm. long or less) or if longer (up to 
9-10 mm. long), then oblong, usually hooded at the tips and loosely connate 
in the bud; inner surface of calyx-lobes glabrous or pubescent; flowers 

Bracteoles early deciduous (just before, or usually much before anthesis) or 

wanting (unknown in E. quebradensis), subulate or linear or sometimes 

lanceolate to ovate, narrowed at base, not connate. 

Inflorescence, especially the hypanthium, heavily pubescent or tomentose 

with gray, white, or light yellowish hairs. 

Branchlets and inflorescence tomentose with crisped, matted hairs; flowers 

sessile in small clusters; leaves lanceolate, 0.8-1.5 cm. wide, 5-6 times as 

long; Lambayeque E. quebradensis McVaugh. 

Branchlets and inflorescence strigose or velutinous, the hairs mostly straight 
and ascending; flowers pedicellate; leaves ovate to elliptic or oblong. 

Leaves 2-6 cm. long; flowers mostly 1-2 pairs, often from the lowest axils 
of new leafy branches. 


30. Bracteoles ovate or lanceolate, blunt; pedicels, hypanthium and calyx 
velutinous; Junln E. barbata McVaugh. 

30. Bracteoles subulate; hypanthium closely strigose, the pedicel sparsely so, the 
calyx glabrate; Bolivia E. mandonii McVaugh. 

29. Leaves 8-24 cm. long; flowers up to 6 pairs in axillary racemes; bracteoles 

subulate; Amazon basin E. prosoneura Berg. 

27. Hairs of the inflorescence red or copper color or dark purplish brown. 

31. Hairs coarse, dark purplish brown; leaves narrowly elliptic, 14-24 cm. long. 

E. gomesiana Berg. 

31. Hairs slender, reddish or pale copper color; leaves broader, mostly less than 
15 cm. long. 

32. Bracts 1 mm. long, broadly rounded at tips and convex on the backs, in- 
conspicuous; calyx-lobes 2.5 mm. long, pubescent except at tips, triangular, 
the long points longer than the corolla in bud E. aerosa McVaugh. 

32. Bracts elongate, 1.5-7 mm. long, delicate, loosely imbricated in 4 ranks 
at the bases of racemes and new shoots; calyx-lobes short and broadly 
rounded, 1.5 mm. long, glabrous both sides, much shorter than the corolla 
in bud E. patrisii Vahl. 

26. Bracteoles persistent, usually until after the fruit falls, lanceolate or broader, 
broad-based, often connate and involucre-like. 

33. Calyx-lobes oblong or elliptic, longer than the petals in bud, hooded and 
thickened at tips, connate in bud below the middle but separating and 
reflexed at anthesis, glabrous within; anthers 1-1.5 mm. long. 

34. Buds 12-15 mm. long; calyx-lobes 8-10 mm. long in flower; stamens 300 
or more E. schunkei McVaugh. 

34. Buds 9 mm. long or less; calyx-lobes 3-6 mm. long in flower (up to 9 mm. 
long in fruit) ; stamens 75-175. 

35. Lateral veins about 20 pairs; buds 6 mm. long, closed at the tip; pedicels 
8-20 mm. long E. hexovulata McVaugh. 

35. Lateral veins 6-15 pairs; calyx-lobes distinct in the bud at least distally; 
pedicels 2-8 mm. long (or a little more in fruit). 

36. Buds 4.5-5 mm. long; calyx-lobes 3-3.5 mm. long; style 6-7 mm. long. 

E. cuspidifolia DC. 

36. Buds 6-9 mm. long; calyx-lobes 3.5-6 mm. long (to 9 mm. in fruit); style 
(unknown in E. acrensis) 10 mm. long. 

37. Calyx permanently velutinous; lateral veins 12-15 pairs, the basal not 
strongly ascending; leaves with numerous small convex glands above. 

E. acrensis McVaugh. 

37. Calyx glabrate; lateral veins 6-10 pairs, the basal ones often strongly ascend- 
ing and not forming a definite marginal vein; leaves eglandular and very 
smooth above E. feijoi Berg. 

33. Calyx-lobes various, not hooded or thickened at the tips or connate below 
the middle; anthers usually less than 1 mm. long. 

38. Inflorescence, including the hypanthium, appressed-pubescent, velutinous 
or tomentose with reddish or reddish-brown hairs. 

39. Bracteoles ovate, 1.5 mm. long, not connate. 

42. Calyx-lobes 3-4 mm. long, concave, dehiscent after anthesis; buds 4 mm. 
long; inflorescence and branchlets closely and finely tomentose with very 
small dark red-brown hairs; leaves 16-24 cm. long, the arcuate-ascending 
lateral veins scarcely forming a marginal vein E. curvivenia McVaugh. 

42. Calyx-lobes 4-6 mm. long, flattened, or cucullate and reflexed or spreading 
after anthesis; hairs appressed-ascending or spreading, lustrous, coppery or 
pale red, up to 0.5 mm. long; marginal veins about as strong as the laterals. 


Leaves 4.5-8 cm. long, with 7-10 pairs of lateral veins; calyx-lobes flat and 
spreading after anthesis, the inner pair truncate, obovate. 

E. crucicalyx McVaugh. 
Leaves 15-21 cm. long, with 12-15 pairs of lateral veins; calyx-lobes reflexed 

and cucullate after anthesis E. acrensis McVaugh. 

Bracteoles broadly ovate to rotund, fused by the basal margins and forming 
an involucre-like cupule beneath the flower. 

Veins impressed above, elevated and conspicuous beneath; leaves pale and 
smooth or appressed-pubescent beneath; inflorescence softly rufous-tomen- 
tose; margins of the bracteoles and calyx-lobes delicate and fracturing even in 

bud E. macrophylla Berg. 

Veins slightly convex on both surfaces, not conspicuous on either one; lower 
leaf surface glistening, irregularly cellular-honeycombed or obscured by tiny 
hairs; inflorescence puberulent with short crisped hairs; bracteoles and calyx 
relatively tough and unbroken even in age. 

Calyx-lobes 6 mm. long; leaves 12-19 cm. long, with 12-15 pairs of lateral 
veins, appearing loosely cellular-honeycombed beneath when viewed with a 

lens E. heterochroma Diels. 

Calyx-lobes 1.5-2 mm. long; leaves 6-13 cm. long, with 6-10 pairs of lateral 
veins, the lower surface obscured by tiny glistening hairs. 

E. versicolor McVaugh. 

Inflorescence, particularly the hypanthium, glabrous, or pubescent with pale 
white or yellowish hairs (hairs of young shoots sometimes reddish). 
Leaves coarsely impressed-punctate and dark above, the lateral veins not 
at all or scarcely apparent; inflorescence usually with some elongate nodes, 
appressed-pubescent with pale hairs; blades 10 cm. long or less. [Including 
E. biflora (L.) DC. and E. inundata DC.] 

Leaves not impressed-punctate, the lateral veins usually apparent above, 
the blades often more than 10 cm. long; nodes of raceme approximate. 
Midvein elevated above in a sharply defined pubescent ridge; bracts 1-3.5 

mm. long, 4-ranked at base of pedicels E. subterminalis DC. 

Midvein impressed to convex above; bracts not 4-ranked, usually incon- 
Flowers usually 2, one from each side of the axillary bud; leaves mostly 3-6 

cm. long, obtuse E. punicifolia (HBK.) DC. 

Flowers fasciculate or in approximate pairs in short racemes, usually 4 or 
more; leaves mostly larger and acuminate. 

Inflorescence, including hypanthium, densely hispidulous with erect pale 
hairs about 0.1 mm. long; raceme up to 8 mm. long, with 2-5 pairs of flowers 
on pedicels 2-4 mm. long; leaves elliptic, 4-7 cm. long, subcaudate-acuminate. 

E. micranthoides McVaugh. 

Inflorescence, at least the hypanthium, glabrous or sparingly strigose; 
leaves and inflorescence various. 

Branchlets, petioles and inflorescence somewhat strigose with white appressed 
hairs; flowers middle-sized, the buds 5-10 mm. long, the disc 3-4 mm. wide; 
leaves 9-18 cm. long, 2.5-3 times as long as wide. 

Buds 7-10 mm. long; marginal vein evident, 2-3 mm. from margin; petiole 
heavily strigose with hairs up to 0.5-0.8 mm. long, spirally longitudinally 
furrowed and transversely wrinkled; hypanthium with a few long hairs. 

E. spruceana Berg. 

Buds 5-6 mm. long; marginal vein not distinct from the laterals; petiole 
nearly glabrous, merely irregularly wrinkled in drying; hypanthium with 
numerous very fine short hairs E. atrosquamata McVaugh. 


48. Branchlets, petioles and inflorescence glabrous, or pubescent with very small 
dibranchiate hairs, minute erect hairs or sparse stiff and ascending reddish 
or pale hairs; flowers smaller, the buds 5 mm. long or less, the disc 1-2.5 
mm. wide. 

50. Leaves elliptic-ovate, 5-6 cm. long or less; midvein flat above; plants sparingly 
strigose with reddish or yellowish hairs up to 0.5 mm. long, the hypanthium 
glabrous; flowers tiny, the calyx-lobes 1 mm. long, the disc 1 mm. wide; 
stamens 36 or fewer E. malpighioides (HBK.) DC. 

50. Leaves usually more than 5 cm. long, rarely ovate; pubescence not as above; 
flowers larger; stamens 60-75 or more. 

51. Flowers sessile or nearly so, in clusters on the stems, the pedicels 2 mm. long 
or less (up to 4 mm. in fruit); styles (as far as known) 10 mm. long or more; 
hypanthium often obconic E. nigra DC. 

51. Flowers on definite slender pedicels mostly 4-10 mm. long; style mostly less 
than 10 mm. long; hypanthium mostly campanulate. 

52. Midvein on upper surface of leaf flat or convex (then sometimes with a shallow 
median furrow near base), never narrowly impressed. [Includes E. egensis 
DC., E. flavescens DC., E. dittocrepis Berg, and E. ochrophloea Diels.] 

52. Midvein impressed on the upper surface, the actual vein narrow and often 
obscured in the bottom of a deep fold or furrow. 

53. Leaves mostly elliptic-lanceolate, long-pointed, very smooth and glabrous 
above; lateral veins inconspicuous, sometimes reddish; plant nearly glabrous, 
the branchlets and pedicels sometimes minutely and sparsely hispidulous. 

E. schomburgkii Benth. 

53. Leaves elliptic to obovate, shortly and often bluntly acuminate, the veins 
usually somewhat elevated and forming an evident reticulum on the upper 
surface; pubescence various. 

54. Inflorescence glabrous or minutely pubescent with pale erect hairs; leaves 
elliptic, blunt-pointed or obscurely acuminate E. tapacumensis Berg. 

54. Inflorescence more or less appressed-puberulent with minute and partly 
dibranchiate hairs. 

55. Leaves oblanceolate or obovate, 14-18 cm. long, coarsely veiny-reticulate. 

E. discrete, McVaugh. 

55. Leaves elliptic, 7-11 cm. long, finely veiny-reticulate. 

E. quadrijuga McVaugh. 

2. (See first "2," p. 194.) Flowers solitary or in 3- to many-flowered dichasia, 
the terminal flower or flowers usually present and sessile; flowers, if solitary, 
arising directly from the axil of a foliage leaf, not from the lower bracteate 
nodes of new branches which are leafy above or from the bracteate nodes of 
abortive axillary buds; bracts and bracteoles mostly linear or lanceolate, 
scarious, deciduous at anthesis or often much before this. 

56. Primary branches of the inflorescence racemosely arranged, i.e., in decussate 
pairs (see first "3," p. 195) E. stipitata McVaugh. 

56. Cymes (dichasia) arising directly from the leaf axils, sometimes irregularly 
branched but the branches not in decussate pairs; bracteoles usually closely 
subtending the flowers. 

57. Flowers 5-merous; leaves coriaceous, yellow-green and vernicose, subsessile, 
orbicular to broadly ovate, 1.5-7.5 cm. long and wide; Pacific slope, Lima. 

E. quinqueloba McVaugh. 

57. Flowers 4-merous; leaves various; species mostly of the inter-Andean valleys. 

58. Flowers solitary (see second "58," p. 201, for plants with flowers partly 
solitary and partly in threes or more numerous). 


59. Leaves 6 mm. long or less, rigidly coriaceous, vernicose, prominently im- 
pressed-punctate on both sides; flowers numerous toward the tips of the 
branches, on peduncles 2-3 mm. long E. minimifolia McVaugh. 

59. Leaves mostly 1-2.5 cm. long, the texture various; blades not impressed- 
punctate beneath, sometimes obscurely so above; flowers relatively few, not 
crowded in the terminal axils, the pedicels (or peduncles) usually much 

60. Hypanthium with strong longitudinal angles; leaves often suborbicular, 
grayish green and closely appressed-pubescent beneath. . .E. oreophila Diels. 

60. Hypanthium not angled; leaves rarely suborbicular, neither grayish green 
nor closely appressed-pubescent. 

61. Hypanthium narrowly obconic, attenuate at base; pedicels 12-20 mm. long; 
calyx-lobes glabrous within; leaves about 2.5 cm. long, mostly acutely 
narrowed to both ends E. myrtomimeta Diels. 

61. Hypanthium narrowly campanulate or broader, blunt or rounded at base 
and usually broader than the markedly compressed pedicel; calyx-lobes 
appressed-pubescent within (sometimes sparingly so in E. myrsinoides with 
pedicels 2-6 mm. long); leaves various. 

62. Leaf margins pale, cartilaginous-thickened and revolute, often irregularly 
roughened and apparently denticulate because of the prominent glands; 
flowers always solitary, the pedicels filiform, little compressed, scarcely 1 mm. 
wide at summit, 10-20 mm. long or more E. cartilaginea McVaugh. 

62. Leaf margins scarcely thickened or roughened, the glands scarcely or not 
at all apparent in mature leaves; peduncle 1- or 3-flowered, usually markedly 
compressed, enlarged distally and 1 mm. wide or more. 

63. Branchlets coarsely appressed-pubescent but the inflorescence glabrous; 
peduncle 1- or 3-flowered, if 1-flowered 5-15 mm. long, compressed, up to 
1.5 mm. wide; stamens about 250; Cuzco E. indifferens McVaugh. 

63. Branchlets from nearly glabrous to pubescent, if coarsely pubescent the 
inflorescence not markedly less so; hypanthium densely strigose; peduncle 
1- or 3-flowered; stamens 50-100. 

64. Leaves mostly less than 1.5 cm. long, obovate, and cuneate at base; flowers 
all solitary, on pedicels mostly 2-6 mm. long; hypanthium mostly glabrous; 
northern Peru and Ecuador E. myrsinoides (HBK.) Diels. 

64. Leaves 1.2-4 cm. long, broadly elliptic to ovate, rounded at base; peduncles 
usually 3-flowered. 

65. Branchlets glabrous or sparingly strigose; Ecuador and Colombia. [Includes 
E. hallii Berg and E. foliosa (HBK.) DC.] 

65. Branchlets densely strigose with straight appressed white or brownish hairs; 
southern Peru (Arequipa) E. ferreyrae McVaugh. 

58. Inflorescence a 3- to many-flowered dichasium, the central flowers usually 
present and sessile (see first "58," p. 200, for plants with solitary flowers 
at least in part). 

66. Veins of the lower leaf-surface raised in a conspicuous reticulum, the veinlets 
enlarged, with appearance of having softened and fused; terminal flowers 
sub-pedicellate E. fimbriata (HBK.) DC. 

66. Veins, if apparent, slender and inconspicuously reticulate; pubescence and 
inflorescence various; terminal flower usually closely sessile. 

67. Inflorescence usually reddish purple, glabrous or essentially so, stout, the 
peduncle 2-3.5 mm. wide near summit; flowers usually 3-7, large, the style 
10-11 mm. long E. rhopaloides (HBK.) DC. 

67. Inflorescence green or brownish in drying, glabrous or variously pubescent; 
peduncle 2 mm. wide at summit or usually less; flowers smaller, the style 
8.5 mm. long or less. 


68. Inflorescence 7-flowered (or the smaller ones on the same plant 3-flowered) 
or repeatedly branched and many-flowered; plants usually markedly strigose 
or otherwise pubescent, especially the inflorescence and hypanthium. 

69. Inflorescence loosely pubescent or tomentose with soft curved or spreading 
hairs; dichasia usually with more than 7 flowers. 

70. Pubescence of whitish spreading hairs 0.5-1 mm. long; leaves mostly obovate, 
5 cm. long or less; northern Peru (Huanuco, Cajamarca). 

E. lindleyana (HBK.) DC. 

70. Pubescence of rufous curved matted hairs; leaves elliptic-ovate, 6.5-11 cm. 
long; Bolivia E. pearcei McVaugh. 

69. Inflorescence thinly appressed-puberulent or strigose with shorter hairs; 
leaves and number of flowers variable. 

71. Leaves less than 2 cm. long, obovate, cuneate; flowers small (style 3.5-4.5 
mm. long), in terminal clusters of 2- to 3-forked cymes; Bolivia. 

E. osteomeloides (Rusby) McVaugh. 

71. Leaves, if less than 2 cm. long, usually elliptic or suborbicular, rarely cuneate; 
style 5-7 mm. long; dichasia axillary, 3- to 7-flowered; central Peru to 

72. Branchlets thinly appressed-puberulent; leaves elliptic-oblong, 2.5-5 cm. 
long; disc 3.5-4 mm. wide, the hairy staminal ring prominent; calyx-lobes 
2.5-3 mm. long; stamens more than 200 E. limbata (HBK.) DC. 

72. Branchlets coarsely pubescent with spreading-ascending hairs; leaves ellip- 
tic to suborbicular, 0.5-2 cm. long; disc 2.5 mm. wide, the staminal ring 
inconspicuous, nearly glabrous; stamens 50-60; central Peru (Ayacucho; 
Apurimac) E. bifurcata McVaugh. 

68. Inflorescence 3-flowered (or if occasionally 7-flowered, glabrous) or the 
flowers solitary in some or all of the axils; hypanthium glabrous to variously 

73. Hypanthium with strong longitudinal angles; leaves often suborbicular, 
grayish green and closely appressed-pubescent beneath; Cuzco. 

E. oreophila Diels. 

73. Hypanthium not angled; leaves various, rarely suborbicular, neither grayish 
green nor closely appressed-pubescent. 

74. Branchlets nearly or quite glabrous, sometimes thinly strigose; northern 
Peru (Libertad) to Colombia. [Includes E. compressa (HBK.) DC., E. 
discolor (HBK.) DC., E. foliosa (HBK.) DC., and E. hallii Berg.] 

74. Branchlets with numerous appressed or spreading stiff hairs up to about 
0.5 mm. long; southern Peru (Arequipa; Cuzco). 

75. Inflorescence glabrous; stamens about 250; Cuzco. .E. indifferens McVaugh. 
75. Inflorescence rather densely strigose; stamens about 50; coastal hills, Are- 
quipa E. ferreyrae McVaugh. 

Eugenia acrensis McVaugh, sp. nov. 

Frutex, gemmis racemisque pilis lucidis cupreis, adpressis obtectis; foliis 15-21 
cm. longis acuminatis, venis utroque latere 12-15; racemis abbreviatis; bracteolis 
persistentibus, ovatis nee connatis, 1.5 mm. longis; calycis lobis intus glabris, in 
alabastro corollam cooperientibus, ad florendi tempus reflexis cucullatisque. 

A shrub 3.5 meters high, nearly glabrous except for the coppery-velutinous 
inflorescence and vegetative buds; leaves elliptic, 15-21 cm. long, 2.5-3 times as 
long as wide, with somewhat impressed veins, the marginal vein about as strong 


as the laterals and arched between them; calyx at anthesis splitting in the sinuses 
down to the level of the staminal ring; stamens about 200; anthers 1.2 mm. long. 

This species, with E. cuspidifolia DC. and E. feijoi Berg, belongs 
to a somewhat ill-defined group characterized by elongate anthers 
and hooded and partially or completely connate calyx-lobes which 
are reflexed and cucullate after anthesis. According to Amshoff 
(Rec. Trav. bot. ne'er!. 42: 21. 1949, and in Pulle, Flora of Suriname 3, 
pt. 2: 58, 105. 1951), these species comprise the genus Catinga Aubl. 
Plants of this affinity are rather abundant in the upper Amazon 
Basin, but the available collections are few, and the taxonomy of the 
group is in need of study and clarification. The degree of coherence 
of the calyx-lobes in the bud varies from almost none in the present 
species to almost complete union in the newly described E. hex- 
ovulata, and in specimens identified by Amshoff as Catinga moschata 

Brazil, Acre: Basin of Rio Purus, near mouth of Rio Macauhan 
(tributary of Rio Yaco), Lat. 9 20' S., Long. 69 W., on terra firma, 
Aug. 22, 1933, B. A. Krukoff 5619 (NY, type). 

Eugenia aerosa McVaugh, sp. nov. 

Arbor, ramulis gemmis paniculisque, pilis pallide cupreis, adpressis obtectis; 
foliis 13-15 cm. longis acuminatis, venis utroque latere 8-10; racemis abbreviatis; 
bracteolis deciduis; calycis lobis deltoideis, 2.5 mm. longis, ut videtur valvatis, ad 
florendi tempus reflexis. 

A tree up to 15 meters high with elliptic-oblong leaves 2.3-3 times as long 
as wide, resembling E. patrisii but lacking the 4-ranked imbricated bracts of that 
species; buds 6 mm. long, pyriform, the long-pointed calyx-lobes exceeding the 
globe of the petals until anthesis; style 10 mm. long; stamens about 100; fruit 
ellipsoid, 2.3 cm. long. 

The single fruiting specimens, probably belonging to this species, 
are from Krukoff's no. 6870, from Livramento, Amazonas, Brazil. 

Peru, Loreto: Mishuyacu, near Iquitos, forest, elev. 100 meters, 
May-June, 1930, G. King 1535 (F 627573, type; NY; US). Iquitos, 
Aug., 1925, G. Tessmann 5355 (fragment, F; G; NY). 

Eugenia atroracemosa McVaugh, sp. nov. 

Arbor, rufo- vel atro-velutina, foliis 11-18 cm. longis acuminatis, venis 
utroque latere circiter 20; racemis 5-7.5 cm. longis validis, floribus 12-20; disco 
3-3.5 mm. lato. 

A tree 4 meters high with oblong, ovate or oblanceolate leaves mostly 2-3 
times as long as wide, readily recognized by the stout reddish brown or almost 
black velutinous racemes; flowers moderately large, the buds 7 mm. long; calyx- 
lobes 2.5-3 mm. long; style 6-6.5 mm. long; stamens 125-150. 


Peru, San Martin: Zepelacio, near Moyobamba, elev. 1,100-1,200 
meters, Oct.-Nov., 1933, G. Klug 3368 (A; F 736442, type; GH; NY; 
US). Brazil, Amazonas: Sao Paulo de Olivenca, Krukoff 8738 (NY). 

Eugenia atrosquamata McVaugh, sp. nov. 

Arbor vel frutex, appresse pubescens, ramulis petiolisque et racemis, pilis 
albidis usque ad 0.2 mm. longis obsitis; foliis 9-15 cm. longis obtuse acuminatis; 
nervo medio supra planiusculo vel concavo; venis utroque latere 6-10, supra 
planiusculis vel convexis; racemis abbreviatis; alabastris 5-6 mm. longis; calycis 
lobis inaequalibus, exterioribus 3.5 mm. longis, 4 mm. latis; bracteolis persistenti- 
bus glabris, ovatis, 1.5 mm. longis et latis. 

A shrub or tree with elliptic leaves about 2.5 times as long as wide, the rela- 
tively few lateral veins ascending but usually not forming a definite marginal vein. 
Flowers 4-6 pairs, middle-sized, with dark contrasting bracts and bracteoles; 
disc 3 mm. wide; style 7 mm. long; stamens about 100. 

This plant is contrasted in the key with E. spruceana Berg, which 
it somewhat resembles, but its actual relationships are obscure 
to me. 

Peru, Loreto: Iquitos, May, 1925, G. Tessmann 5130 (G, type; 
NY). Colombia, Amazonas: Trapecio amazonico, Loretoyacu 
River, elev. 100 meters, March, 1946, R. E. Schultes 7133 (US). 
Univ. of Mich. Neg. 439. 

Eugenia barbata McVaugh, sp. nov. 

Ut videtur frutex, ramulis hispidulis, floribus cinereo-flavido-velutinis; foliis 
3-5 cm. longis, acutis vel obtusis, venis inconspicuis; racemis abbreviatis, floribus 
plerumque 2, interdum 4; bracteolis deciduis, obtusis; alabastris 9-10 mm. longis 

Probably a shrub, 4-8 meters high with coriaceous broadly elliptic leaves and 
markedly hairy, rather large flowers in small nearly sessile clusters or solitary 
at the lower nodes; calyx-lobes suborbicular, 4-5 mm. long; style 9-10 mm. long; 
stamens about 250. 

This plant superficially resembles E. pungens Berg, a species 
which ranges from the lowlands of Bolivia to Uruguay; in E. pungens, 
however, the leaves are normally aristate, and the pubescence is 
nearly white. 

Peru, Junin: Valle del Rio San Bernardo, elev. 2,200-2,300 
meters, en monte bajo, abierto, compuesto de arbustos, Apr. 2, 
1913, A. Weberbauer 6558 (F; GH; US 1497251 pro parte, type; 
USM). Pariahuanca, elev. 1,800 meters, bosque compuesto de 
arboles y arbustos, Apr. 8, 1913, Weberbauer 6594 (F; GH; US; 


Eugenia bifurcata McVaugh, sp. nov. 

Frutex parvifolius, appresse pubescens vel strigosus; foliis subrotundatis, 
0.5-2 cm. longis; floribus subcorymbosis, dichasiis 3-7-floris; stylo 5-7 mm. 
longo; disco 2.5 mm. lato, fere glabro; staminibus 50-60. 

A shrub to 3 meters high, with elliptic to suborbicular, ovate or obovate 
leaves up to 1.5 times as long as wide, the blades coriaceous and obscurely veined. 
The rather slender, several-flowered, corymbosely aggregated dichasia 2-3 cm. 
long are distinctive. Flowers small, the buds 3-4 mm. long. 

This species is superficially similar in flower and inflorescence 
characters to E. osteomeloides (Rusby) McVaugh, a Bolivian species, 
but differs in a number of individually trivial respects, the most 
notable of which are contrasted above in the key to species. 

Peru, Apurimac: Valley of the Rio Pampas, Lat. 13 20' S. to 
13 30' S., elev. 2,600 meters, A. Weberbauer 5842 (F; GH; US 
1497208, type) ; Pincos, in rain-green shrubland, elev. 2,700 meters, 
Feb. 19, 1939, Stork & Horton 10678 (F). Ayacucho: Hills from 
River Pampas to Ocros, elev. 9,000-10,000 feet, R. Pearce (BM). 

Eugenia calva McVaugh, sp. nov. 

Arbor, appresse pubescens, glabrescens, floribus glabris; foliis 9-15 cm. longis 
acuminatis, venis utroque latere 12-15, nervo marginali aperto; racemis 4.5-5.5 
cm. longis; alabastris 2-3 mm. longis glabris, calycis lobis fimbriatis, intus pu- 

A tree 6 meters high with lance-ovate or elliptic leaves 2.5-3 times as long 
as wide; because of the glabrous flowers this species is not likely to be confused 
with any other except E. florida DC., from which it differs in the characters set 
forth in the key to species. Flowers small, the larger calyx-lobes 1.5 mm. long; 
stamens about 100. 

Peru, Loreto: Florida, Rio Putumayo, mouth of Rio Zubineta, 
elev. 180 meters, forest, May-July, 1931, G. King 2146 (F 668811, 
type; G; GH; NY; US). 

Eugenia cartilaginea McVaugh, sp. nov. 

Frutex vel arbor parvifolius, appresse hispidulus vel gemmis strigosis; foliis 
1-2.5 cm. longis, coriaceis, obtusis vel acutis; marginibus pallidis, incrassatis, 
paulum revolutis, glandulosis; floribus solitariis, axillaribus, vel infimis ex axillis 
bractearum caducarum oriundis; hypanthio obconico; pedicellis filiformibus 
10-20 mm. longis; calycis lobis intus pubescentibus. 

A shrub or small tree 4-6 meters high with lustrous ovate or elliptic leaves 
1.5-2 times as long as wide, the margins often irregularly roughened by the 
prominent glands; bracteoles deciduous; flowers rather small, the disc 2.5-3 mm. 
wide; style 5-6 mm. long; stamens about 75. 

This species evidently belongs with the complex which includes 
some species with 3-flowered dichasia, e.g., E. discolor (HBK.) DC. 


It is similar on the one hand to E. orthostemon Berg, a species which 
ranges from central Ecuador to Colombia. In E. orthostemon, 
however, pubescence is of more general occurrence on the flowers 
and pedicels, and the leaves are larger (up to 3.5 cm. wide and 6-7 
cm. long) ; the leaves also lack the cartilaginous margins, prominent 
glands and mucronate tips which characterize those of E. carti- 
laginea. The latter is similar in many respects, on the other hand, 
to E. myrsinoides (HBK.) Diels, from which it may be distinguished 
by the characters given in the key. 

Peru, Junin: Valley of the Rio Mantaro near Huachicna, elev. 
2,300 meters, Weberbauer 6548 (F; GH; NY; US; USM). Huan- 
cavelica: Pampas-Salcabamba trail, elev. 2,500 meters, Stork & 
Horton 10443 (F 1180795, type). 

Eugenia chartacea McVaugh, sp. nov. 

Arbor, glabra (disco sparse piloso; calycis lobis petalisque minute ciliatis); 
foliis late ellipticis vel ovatis, 11-21 cm. longis obtuse acuminatis; venis 7-8, supra 
convexis sed sulcatis; nervo marginal! aperto; petiolo 2 mm. crasso canaliculate; 
racemis abbreviatis; alabastris 15-18 mm. longis; calycis lobis rigidis, chartaceis, 
imbricatis, 10-12 "mm. longis; antheris 2-2.3 mm. longis. 

A tree 15 meters high with very broad leaves about 1.5 times as long as wide, 
relatively few veins, and small clusters of large flowers produced on old wood. 
Disc quadrangular, 4-5 mm. wide; style probably 15 mm. long or more; stamens 

Brazil, Amazonas: Basin of Rio Jurua, on high terra firma, near 
mouth of Rio Embira (tributary of Rio Tarauaca), Lat. 7 30' S., 
Long. 70 15' W., June 21, 1933, B. A. Krukoff 4951 (NY, type; US). 

Eugenia crucicalyx McVaugh, sp. nov. 

Frutex, ramulis gemmisque et racemis pilis appressis, pallide rufulis, obtectis; 
foliis 4.5-8 cm. longis brevi-acuminatis; venis utroque latere 7-10, utrinque 
elevatis; racemis abbreviatis; hypanthio paulum angulato, rufo-velutino; calycis 
lobis 5-6 mm. longis, quam hypanthio 2-3-plo longioribus, intus glabris, ad 
florendi tempus patentibus; bracteolis 1.5 mm. longis persistentibus, ovatis, 
distinctis; stylo 8 mm. longo. 

A shrub with concolorous, lustrous elliptic leaves about twice as long as wide 
and a well-defined marginal vein; buds 7-8 mm. long; calyx-lobes in unequal 
pairs, the inner ones obovate and truncate, the outer ovate-triangular, bluntly 
pointed; disc 2-2.5 mm. wide; stamens 200-250. 

Peru, San Martin: Tarapoto, elev. 360-900 meters, forest, Dec. 4, 
1929, L. Williams 6216 (F 627015, type). Univ. of Mich. Neg. 466. 

Eugenia curvipilosa McVaugh, sp. nov. 

Frutex, pallide brunneo-pilosus; foliis 4-7 cm. longis breve lateque acuminatis, 
supra sub-vernicosis scabriusculisque, subtus tomentosis; racemis 1-2 cm. longis 


tenuibus; floribus 6-12, divaricatis, pedicellis 4-6 mm. longis; hypanthio strigoso, 
disco 1.5-2 mm. lato. 

A shrub 3 meters high, with subcoriaceous elliptic-ovate leaves 1.5-2 times 
as long as wide, the midvein and the about 10 lateral veins somewhat impressed 
above; bracteoles persistent, connate, glabrescent, not strongly glandular; flowers 
small, the buds 4 mm. long, style about 4(?) mm. long, stamens 50-60. 

Superficially suggesting E. limbosa Berg, but readily differenti- 
ated from that species by the leaf and pubescence characters used 
in the key. 

Peru, Cajamarca: Valley of the Rio Llaucan, near Pion, elev. 
1,700-1,800 meters, June, 1915, A. Weberbauer 7137 (F 628083, 
type; GH). Univ. of Mich. Neg. 447. 

Eugenia curvivenia McVaugh, sp. nov. 

Frutex, minute fusco-rufo-tomentosus; foliis 16-24 cm. longis acuminatis 
mox glabrescentibus; venis utroque latere circiter 10, impressis, arcuatim adscen- 
dentibus, apicem versus sensim extenuatis; nervo marginali nullo; racemis ab- 
breviatis; bracteolis 0.8-1.5 mm. longis persistentibus, non connatis; floribus 
minusculis, alabastris 4 mm. longis, calycis lobis 3-4 mm. longis distinctis, deciduis; 
disco 2 mm. lato; stylo 5 mm. longo. 

A shrub 6 meters high with large, broadly elliptic-ovate coriaceous veiny 
leaves 2-2.5 times as long as wide, stout dark petioles, and (in proportion to the 
leaves) very small flowers in small clusters in the axils; stamens 50-60. 

Brazil, Amazonas: Basin of Rio Jurua, near mouth of Rio 
Embira, tributary of Rio Tarauaca, 7 30' S. Lat., 70 15' W. Long., 
June 27, 1933, B. A. Krukoff 5045 (NY, type; US). 

Eugenia dibranchiata McVaugh, sp. nov. 

Arbor, innovationibus petiolisque et racemis, pilis rufis appressis, partim 
dibranchiatis, obsitis; foliis glabris vel appresse strigosis, 8.5-14 cm. longis anguste 
acuminatis; nervo medio supra elevato; venis utroque latere 10-15; nervo margin- 
ali interrupto, exiguo; racemis 3-6 cm. longis tenuibus; bracteolis obtusis, con- 
natis, strigosis vel ciliatis, nee glandulosis; disco 2.5-3 mm. lato; stylo 7 mm. 

A tall forest tree with inconspicuously veined elliptic leaves 2-3 times as long 
as wide, and slender axillary racemes with 4-7 pairs of flowers; hypanthium 
pubescent without; calyx-lobes about 2 mm. long, pubescent on both sides; 
stamens 60-75, up to 10 mm. long. 

This species is apparently most closely related to E. florida DC.; 
the leaves suggest in color and texture those of E. florida, but the 
marginal vein is more uniformly developed in E. dibranchiata, and 
the midvein is elevated, whereas in E. florida it is flat or somewhat 
impressed. Specimens suggesting intermediates between these two 
species have been collected near Palmares, Brazil, not far from the 


Peruvian border (Krukoff no. 8322), and also at Zepelacio, near 
Moyobamba, Peru (Klug no. 3304). These plants have rather 
well-developed marginal veins, and the pubescence and floral charac- 
ters are nearly as in E. dibranchiata (except that the hypanthium is 
glabrous in Klug's collection), but the midveins are impressed as in 
E. florida and the young leaves are pubescent along the midvein 
as in that species. It appears also that E. dibranchiata is related to 
the species complex which includes E. riparia DC., but differs in the 
thin and concolorous leaves, poorly developed marginal vein, ele- 
vated midrib, and the shorter and more compact racemes with smaller 
flowers, relatively longer styles and a far greater proportion of 
branched hairs. 

Peru, Loreto: Caballo-Cocha, Aug. 13, 1929, L. Williams 2449 
(F 615017, type); Williams 2415 (F), 2416 (F). La Victoria, 
Williams 2934 (F). Leticia, Williams 3145 (F). 

Eugenia discreta McVaugh, sp. nov. 

Arbor, puberulenta, pilis appressis minimis, magna pro parte dibranchiatis, 
obsita; foliis oblanceolatis obovatisve, 14-18 cm. longis acuminatis; venis utroque 
latere 10-15, supra elevatis; nervo medio supra impresso; venulis reticulatis; 
racemis abbreviatis; floribus 4-6 minusculis pedicellatis, alabastris 5 mm. longis; 
bracteolis ovatis obtusis, 2.5 mm. longis persistentibus connatis; disco 2 mm. 
lato; staminibus circiter 75. 

A tree to 11 meters high with reticulate- veiny leaves 2.5-3 times as long as 
wide. The leaves tend to become liver-colored beneath in drying but to remain 
olive-green above. The calyx-lobes are unequal, 1-2.5 mm. long, the style 7-9 
mm. long; fruit globose, said to be yellow. 

Peru, Loreto: Gamitanacocha, Rio Mazan, on river bank, Jan. 
15, 1935, J. M. Schunke 40 (A; F; US 1458946, type). Brazil, 
Amazonas: Near mouth of Rio Embira, Krukoff 4884 (US). 

The Peruvian species related to the above are poorly represented 
in herbaria and are consequently difficult to interpret taxonomically. 
The group may be characterized as having small, slender-pedicellate, 
glabrous flowers in axillary "glomerules" or "fascicles" (actually 
very short racemes), nearly glabrous foliage, the midvein impressed 
above, and the marginal vein relatively far from the margin with 
a distinct sub-marginal vein beyond it, the fruit globose and 1 cm. 
in diameter or less. The species in question include E. schom- 
burgkii Benth., E. tapacumensis Berg, E. maculata Berg, and two 
newly proposed species, E. discreta and E. quadrijuga. A species of 
Amazonian Brazil, E. agathopoda Diels (Verh. Bot. Ver. Brandenb. 
48: 192. 1907), is similar but has larger flowers, longer racemes and 
longer pedicels. 


Because of the small amount of material available for study, no 
really workable key to the above species can be constructed. The 
flowers of all the species are so much alike that they provide no good 
diagnostic characters, and the fruits are almost unknown. In the 
present treatment Eugenia schomburgkii includes chiefly glabrous or 
nearly glabrous plants with lanceolate and long-pointed leaves which 
are very smooth on the upper surface; E. maculata is distinguished 
from broad-leaved extremes of E. schomburgkii, and from the other 
species involved, by its relatively well-developed raceme axis which 
may be as much as 5 mm. long. Eugenia tapacumensis is a plant 
with subcoriaceous, elliptic, bluntly pointed leaves and rather fine, 
inconspicuous veins; E. discreta, known only from the type, has 
mostly oblanceolate and rather large, reticulate-veiny leaves, and 
sparse, appressed dibranchiate hairs in the inflorescence. The 
remaining species, E. quadrijuga, is described from four collections 
which may represent one species or conceivably as many as four. 
The collections agree very well in most features, but differ in leaf 
shape from long-elliptic (as in Ule's no. 9661, which suggests E. 
schomburgkii) to broadly elliptic (as in Klug's no. 3153, which 
suggests E. tapacumensis). The pubescence differs slightly from 
one of the four collections to another, but all have in common the 
minutely bristly ascending hairs of the pedicels. 

Eugenia ferreyrae McVaugh, sp. nov. 

Frutex, ramulis dichasiisque, et innovationibus, insigniter strigosis; foliis 
1-2 cm. longis coriaceis, obtusis vel rotundatis vel emarginatis; nervo medio 
impresso, venis et nervo marginali inconspicuis; floribus solitariis, vel plerumque 
pedunculis 3-floris, 7-10 mm. longis, compressis, apicem versus 1 mm. latis; 
hypanthio ut videtur campanulato, strigoso; calycis lobis intus strigosis vel 
demum glabrescentibus; staminibus ut videtur circiter 50; embryone cotylis 
carnosis, discretis. 

A shrub with stiff smallish, broadly elliptic, ovate or obovate leaves 1.2-1.9 
times as long as wide, sometimes drying bluish- or grayish-green above and dark 
reddish-brown beneath, the upper surface sparingly impressed-puncticulate; 
calyx-lobes broad and coriaceous, 2-3 mm. long; disc 3 mm. wide; style 7-9 mm. 
long; fruit probably globose, about 7-10 mm. in diameter, the seeds 1 or 2, about 
7 mm. long. 

This species is of unusually great interest, as coming from an 
arid area of southern cis- Andean Peru, where no other native species 
of Myrtaceae is known to occur. Unfortunately it is known but 
from two collections, and one fruit only has been available for study. 
The plant resembles in general habit and leaf morphology some of 
the Chilean Myrtaceae, e.g., Reichea coquimbensis (Barn.) Kausel, 


but that species has pentamerous flowers and a very short erect 
radicle, whereas in Eugenia ferreyrae the radicle is accumbent, 
tapering-cylindrical, about 0.5 mm. thick at base, and 2.5 mm. long. 
The embryo of E. ferreyrae, with its separate cotyledons and prom- 
inent radicle, is anomalous in the genus Eugenia, sens, str., but 
apparently agrees well with those of other species of the genus 
Anamomis Griseb. Until a more general survey of fruit and seed 
characters in Anamomis can be made, however, I prefer to refer the 
present species and its close relatives to Eugenia. 

Peru, Arequipa: Prov. Caraveli, Lomas de Chaparra, elev. 500- 
560 meters, Oct. 19, 1946, Ramdn Ferreyra 1483 (USM, type); 
southeast of the port of Chala, highway to Chaparra, elev. 600-750 
meters, Oct. 10, 1955, Ferreyra 11450 (MICH). Univ. of Mich. 
Neg. 427. 

Eugenia hexovulata McVaugh, sp. nov. 

Arbor, appresse pubescens, ramulis petiolis umbellisque, et folii pagina inferiore, 
pilis densis rufis dibranchiatis obtectis; foliis 12-24 cm. longis acuminatis; nervo 
medio supra elevato; venis utroque latere circiter 20, utrinque paulum elevatis; 
nervo marginali aperto; racemis abbreviatis, umbelliformibus, usque ad 15-floris; 
alabastris 6 mm. longis clausis, obtusis vel obscure apiculatis, hypanthio 8- 
angulato; calyce ad florendi tempus usque ad verticem germinis in lobos 4 sub- 
aequales longitudinaliter direpto; lobis ovatis, acutis, 5 mm. longis, intus glabris; 
bracteolis persistentibus, appressis. 

A tree 15 meters high, with lance-oblong leaves about 3 times as long as wide; 
disc 2 mm. wide, the staminal ring extending nearly to the depressed center; 
style 6-7 mm. long; stamens 100-125, the anthers linear, 1-1.2 mm. long; ovary 
bilocular, the ovules 3 in each locule, collateral. 

A plant of uncertain relationships; the small number of ovules 
is unusual but not unknown in the Eugeniinae. An affinity to 
E. feijoi Berg, E. cuspidifolia DC. and their relatives (the supposed 
genus Catinga Aubl.) is suggested by the characters of inflorescence, 
bracteoles and pubescence, the angled hypanthium, the long narrow 
anthers and the broad staminal ring. The venation in E. hex- 
ovulata, however, is anomalous in Catinga, and the closed buds of the 
present species appear to be nearly unique. 

Peru, Loreto : Pumayacu, between Balsapuerto and Moyobamba, 
elev. 600-1,200 meters, forest, Aug.-Sept., 1933, G. King 3194 
(F 715574, type; NY). Univ. of Mich. Neg. 462. 

Eugenia illepida McVaugh, sp. nov. 

Arbor, glabra (disco piloso; bracteolis, petalis et calycis lobis ciliatis); foliis 
4-7.5 cm. latis, 12-26 cm. longis, leviter acuminatis; petiolo 1-2 mm. crasso; venis 


utroque latere 10-15, utrinque paulum elevatis; nervo marginali aperto; racemis 
abbreviates; alabastris 12 mm. longis, supra hypanthium globosis; calycis lobis 
rigidis, glandulosis, late imbricatis, inaequalibus; lobis interioribus 8 mm. longis, 
14 mm. latis; bracteolis persistentibus nee connatis; disco 6-7 mm. lato; antheris 
1-1.3 mm. longis. 

A tree to 20 meters high, with relatively thin oblanceolate or elliptic leaves 
3-3.5 times as long as wide. Flowers in axillary clusters, on pedicels up to 15 mm. 
long; flowers large, the style 16-17 mm. long, the stamens probably about 300. 

Brazil, Acre: Near mouth of Rio Macauhan (tributary of Rio 
Yaco), Lat. 9 20' S., Long. 69 W., on terra firma, Aug. 13, 1933, 
B. A. Krukoff 5482 (NY; US), Aug. 27, 1933, Krukoff 5675 (NY; 
US 1662164, type). Krukoff 's no. 5482 is represented by unbroken 
specimens, with the inflorescence in place; in these specimens, the 
anthers are invariably abnormal because of the attacks of gall- 
forming insects, and no. 5675, with apparently normal flowers, is 
designated as type. 

Eugenia indifferens McVaugh, sp. nov. 

Arbor vel frutex, ramulis foliisque novellis pubescentibus vel strigosis, pe- 
dunculo hypanthioque glabris; foliis rigide coriaceis 1.3-3 cm. longis obtusis, 
maturitate eglandulosis; venis inconspicuis; pedunculis 1-3-floris compressis, 
5-15 mm. longis, apicem versus usque ad 1.5 mm. latis; hypanthio late turbinato; 
calycis lobis intus sericeis; staminibus circiter 250. 

A tree or shrub with ovate, obscurely veined leaves as broad as long, or up 
to twice as long as broad; most of the flowers solitary; buds 3-5 mm. long; disc 3.5 
mm. wide, glabrous, the style 6.5-7.5 mm. long. 

Peru, Cuzco: Urubamba Valley, Hacienda Fanccac, elev. 2,760 
meters, Sept. 10, 1928, F. L, Herrera 2099 (F; US 1422430, type). 
Univ. of Mich. Neg. 442. 

Eugenia longicuspis McVaugh, sp. nov. 

Arbor, ramulis racemisque pilis appresse-adscendentibus obtectis; foliis 18-21 
cm. longis, cuspidato-acuminatis; nervo medio supra impresso; venis utroque 
latere 8-12, supra impressis, subtus strigosis; nervo marginali aperto, supra 
impresso; racemis 6-9 cm. longis validis; disco 6 mm. lato. 

A tree 9 meters high with obovate-oblong leaves about 3 times as long as 
wide, readily recognized by the caudate tips 2-3 cm. long, and the prominent 
pattern formed on the lower surface by the lateral and marginal veins. Flowers 
3-5 pairs, large for this group of species; bracteoles connate, persistent, together 
5 mm. wide; style not seen; stamens more than 200. 

The only collection surely referable to this species is the type, 
which bears flowers from which the corolla, androecium and style 
have fallen. A specimen which seems intermediate between this 
and E. atroracemosa McVaugh is Williams' no. 4747, from Yurima- 


guas, Peru, which bears immature buds and one raceme with half- 
grown fruits. This has the stout, short-pedicelled racemes, short- 
acuminate leaves and non-connate bracteoles of E. atroracemosa, but 
the disc (6-7 mm. wide and glabrous at the center), stamens (about 
200) and style (9 mm. long) of E. longicuspis. The ovules are about 
7 in each locule, and collateral. This may represent an undescribed 
species, but the material is imperfect and is referred with doubt 
to E. longicuspis. A second collection from near Iquitos, Williams' 
no. 8018, is perhaps also referable to this species, but the leaves tend 
to be elliptic and are not markedly caudate-acuminate, the upper 
surface is impressed-puncticulate, the pedicels are 5 mm. long and 
the disc is 4.5 mm. wide; the fruit, nearly at maturity, is subglobose, 
about 1.5 cm. in diameter. 

Peru, Loreto: Mishuyacu, near Iquitos, elev. 100 meters, forest, 
Feb.-Mar., 1930, G. King 855 (F; NY; US 1455842, type). 

Eugenia macrocalyx (Rusby) McVaugh, comb. nov. Caly- 
corectes macrocalyx Rusby, Mem. N. Y. Bot. Gard. 7: 313. 1927. 

A distinctive species, but quite out of place in Calycorectes, 
which was described as having the calyx closed in bud, then splitting 
longitudinally. The present species belongs rather to the genus 
Phyllocalyx Berg, which I believe is better regarded as a part of the 
inclusive genus Eugenia. 

Eugenia mandonii McVaugh, sp. nov. 

Arbor vel frutex, ramulis floribusque et praesertim hypanthio strigosis; foliis 
2-4 cm. longis obtusis acutisve vel subacuminatis; venis utroque latere 5-8 
inconspicuis; pedunculis solitariis, 1-floris, oppositis, ex infimis nodis 2 ramulorum 
hornotinorum oriundis, vel rhachi raro abortiva, abbreviata, efoliata, 1-4-flora; 
bracteolis subulatis 2 mm. longis strigosis, deciduis; calycis lobis rotundatis 
2.5 mm. longis, intus strigosis. 

A tree or shrub with small coriaceous elliptic-ovate leaves 2-3 times as long 
as wide, strigose and vernicose above, with impressed midvein; pedicels 8-16 mm. 
long, subtended by small foliaceous bracts or much smaller scarious bracts 1.5-2 
mm. long; buds 6 mm. long, obovate; disc 4-4.5 mm. wide; style 6.5-9 mm. long; 
stamens about 200. 

Bolivia: Prov. Larecaja, viciniis Ananea, in nemoribus, June, 
1856, G. Mandon 634 (G, type; NY). Univ. of Mich. Neg. 438. 

Eugenia micranthoides McVaugh, sp. nov. 

Arbor vel frutex, minute pubescens; ramulis racemisque dense pilis erectis 
0.1 mm. longis obsitis; foliis ellipticis 4-7 cm. longis subcaudatis supra eglandu- 
losis; nervo medio supra sulcato; venis utroque latere circiter 10, utrinque obscuris; 


racemis brevis, 2-8 mm. longis tenuibus, 4-10-floris; pedicellis 2-4 mm. longis; 
bracteolis persistentibus; calycis lobis 1 mm. longis. 

A shrub or small tree with tiny racemes, small flowers and small narrow leaves 
2-3 times as long as wide, their narrowly acuminate tips 1-1.5 cm. long; disc 
about 2 mm. wide; style 4.5-5 mm. long; stamens about 50. 

This plant is strikingly similar in characters of pubescence and 
in morphology of the inflorescence to E. micrantha (HBK.) DC., but 
in that species the leaves are narrower, prominently veined and 
glandular, the midvein is not impressed above, the flowers are 
smaller and the bracts are persistent. The type locality of E. 
micrantha is near Honda, in the valley of the Rio Magdalena, 
Colombia. A note at the end of the original description, however 
(HBK. Nov. Gen. & Sp. 6: 145 [folio ed. 115]. 1823), reads: "Bon- 
plandius haec specimina in Peru via lecta esse memorat." Ap- 
parently this reference to Peru is erroneous, for the species is well 
known in Colombia but has not been found subsequently in Peru. 

Peru, Lore to: Yurimaguas, Parana Pura, in forest, Oct.-Nov., 
1929, L. Williams 3825 (F), 4102 (F), 4620 (F 623009, type). 

Eugenia minimifolia McVaugh, sp. nov. 

Frutex multiramosus compactus, appresse strigosus; foliis 2.5-6 mm. longis 
vernicosis rigide coriaceis, obtusis, subaveniis, utrinque impresso-punctatis; pe- 
dunculis 1-floris solitariis, 2-3 mm. longis. 

A much-branched sclerophyllous shrub 2 meters high, with the numerous 
white flowers abundant toward the tips of the branches; leaves obovate or elliptic, 
1.5-2 times as long as wide; buds 3.5 mm. long; calyx-lobes broad, 1 mm. long, 
glabrous both sides; style 5.5-6 mm. long; stamens about 125. 

Peru, Huancavelica : Prov. Tayacaja, hills to the left of the Rio 
Mantaro, above the bridge of Chiquiac, low open mountain sides 
covered with evergreen shrubs, elev. 2,600-2,700 meters, Mar. 15, 
1913, A. Weberbauer 6500 (F; GH; US 1497222, type; USM). 

Eugenia multirimosa McVaugh, sp. nov. 

Frutex vel arbor, glomerulis (racemis abbreviatis) puberulis; ramis rufidis, 
exophloeo longitudinaliter rumpente et secedente; foliis subsessilibus, basi cordato- 
auriculatis, plerumque oblanceolatis, ut videtur 25-35 cm. longis, 3-5-plo longior- 
ibus quam latioribus; nervo medio supra impresso; venis utroque latere circiter 
15; petiolo rimoso 3-5 mm. longo, ut videtur 3-5 mm. crasso (exophloeo suberoso 
soluto); racemis abbreviatis; bracteolis persistentibus 1.5 mm. longis; calycis 
lobis suborbiculatis, majoribus 4-5 mm. longis; disco 3.5-4.5 mm. lato. 

A shrub or small tree, well marked by the nearly sessile, cordate-auriculate 
and oblanceolate leaves, and by the tendency for the outer cortical layers of the 
petioles and pedicels to separate irregularly; flowers about 6, on stout pedicels 
8-11 mm. long, not large for the size of the leaves, the hypanthium in anthesis 


3 mm. long; stamens probably about 300; fruit said to be orange, ellipsoid, about 
3 cm. long. 

Peru, Loreto: Yurimaguas, along lower Rio Huallaga, elev. 135 
meters, dense forest, Aug. 23-Sept. 7, 1929, Kittip & Smith 29019 
(NY; US 1462448, type); Yurimaguas, Kittip & Smith 28020 (F; 
NY); Timbuchi, L. Williams 1017 (F). 

Eugenia osteomeloides (Rusby) McVaugh, comb. nov. Myr- 
tus osteomeloides Rusby, Mem. Torr. Bot. Club 6: 36. 1896. M. 
myrciopsis 0. Ktze. Rev. Gen. 3, pt. 2: 92. 1898. E. myrciopsis 
(O. Ktze.) K. Schum. in Just, Bot. Jahresb. 26, pt. 1: 359. 1900. 

This is a Eugenia of the affinity of Anamomis Griseb., not 
a member of the Pimentinae. The cotyledons are large, distinct and 
plano-convex; the radicle is cylindrical and about a third as long as 
the seed. The type specimens of Myrtus myrciopsis, from Santa 
Cruz, Bolivia, are somewhat more densely pubescent than other 
plants from the same area. Possibly the number of hairs is related 
to the fact that Kuntze's plants are in young bud; there do not 
appear to be other significant differences between M. myrciopsis and 
M. osteomeloides. 

Eugenia pearcei McVaugh, sp. nov. 

Frutex sempervirens, rufo-tomentosus; foliis coriaceis, 6.5-11 cm. longis 
obtusis, subtus tomentosis; venis utroque latere circiter 10, arcuatim adscenden- 
tibus, apicem versus sensim extenuatis, venulis nee incrassatis; dichasio plerumque 
7-15-floro, floribus in dichotomis sessilibus; pedunculo 2.5-5 cm. longo, compresso, 
apicem versus usque ad 2 mm. lato; alabastris 4-5 mm. longis; calycis lobis utrin- 
que tomentosis; stylo 5 mm. longo vel longiore; disco 3-4 mm. lato. 

An evergreen shrub 2-3 meters high, with lustrous thick elliptic-ovate leaves 
about twice as long as wide, impressed-punctate above; flowers middle-sized, the 
calyx-lobes 3 mm. wide, 1.5-2.5 mm. long; stamens 200-250; petals lightly tomen- 
tose without. 

A very distinctive species, but unfortunately the only known 
specimen has been severely damaged by insects. 

Bolivia: Hills near Pata, 6,000 feet, Dec., 1864, R. Pearce (BM). 
Univ. of Mich. Neg. 483. 

Eugenia percincta McVaugh, sp. nov. 

Frutex vel arbor, dense fulvo-velutinus; foliis elliptico-oblongis 12-14 cm. 
longis, 6-8.5 cm. latis deltoideo-acuminatis, crassimarginatis; racemis perbrevibus, 
rhachi subnulla et floribus 4 approximatis; calycis lobis 4-5 mm. longis latisve 
concavis, imbricatis, utrinque velutinis; bracteolis deciduis; stylo 14-16 mm. 


A tree or shrub, well marked by the heavy covering of tawny hairs, the 
strong cartilaginous margins of the leaves, and the conspicuous glomerules of 
rather large flowers; disc 4-4.5 mm. wide; stamens about 150. 

Brazil, Guapore": Falls of Madeira, Oct., 1886, H. H. Rusby 2084 
(NY; US 1416665, type). Univ. of Mich. Neg. 453. 

Eugenia percrenata McVaugh, sp. nov. 

Frutex, subglaber; foliis 6-9 cm. longis caudatis, marginibus conspicue 
crenatis; nervo medio supra planiusculo vel elevato; racemis 4-10-floris ab- 
breviatis; bracteolis persistentibus; calycis lobis suborbiculatis vel oblongis, 1.5-2 
mm. longis; disco 1.5 mm. lato. 

A shrub 5 meters high, glabrous except for a few strigose hairs on the bracts 
and on vegetative buds; leaves elliptic, 2.5-3.5 cm. wide, and 3-3.5 times as long 
as wide, the apex caudate-acuminate, the base cuneate and attenuate to a petiole 
1-1.5 mm. thick and 3-4 mm. long; margins deeply crenate with 6-10 notches on 
each side, the notches 1 mm. deep with a gland 0.5-0.8 mm. wide at the base of 
each; midvein flat, or centrally and sharply keeled above, pale and elevated 
beneath; lateral veins about 12 pairs, obscure above, somewhat raised beneath; 
marginal veins about equaling the laterals and arched between them, 1-2 mm. from 
margin; blades when dry dull and dark in color and obscurely gland-dotted above, 
brown and conspicuously glandular beneath; inflorescence a short axillary raceme, 
the axis 2-5 mm. long, with 2-5 approximate decussate pairs of flowers; bracts 
1 mm. long, ovate, ciliate, rounded on the backs; pedicels 5-10 mm. long, some- 
what compressed distally and there 1 mm. wide; bracteoles 0.7 mm. long, per- 
sistent and nearly erect, ovate but trough-shaped and so appearing lance-ovate, 
acute; hypanthium 2 mm. long, constricted at base above the bracteoles; calyx- 
lobes suborbicular or oblong, rounded at apex, strongly reflexed after anthesis, 
1.5-2 mm. long and wide; disc about 1.5 mm. wide, somewhat quadrangular; 
style 3.5 mm. long; stamens about 50, up to 3 mm. long; anthers 0.8 mm. long; 
petals ovate, 3 mm. wide, 3.5 mm. long, yellowish white (Krukoff ) ; ovary bilocular, 
the ovules about 15 in each locule, radiating from a centrally affixed placenta. 

Brazil, Mato Grosso: Near Tabajara, upper Machado River 
region, in terra firma, Nov. 18, 1931, B. A. Krukoff 1368 (MICH, 
type; NY; US). 

This species resembles in habit and in morphology of leaf and 
inflorescence both Eugenia egensis and E. flavescens, but it is readily 
distinguished from these and from other known species by the 
caudate-acuminate and markedly crenate leaves, and by the reflexed 

Eugenia pustulescens McVaugh, sp. nov. 

Arbor vel frutex, plusminusve dense strigosus, pilis ferrugineo-rufidis ap- 
pressis partim dibranchiatis, etiam pilis erectis brevioribus immixtis, obsitus; 
foliis 7-9 cm. longis brevi-acuminatis; racemis usque ad 6 cm. longis validis; 
bracteis bracteolisque et calycis lobis vix pubescentibus, sed glandulis fuscis 
convexis crebro obsitis; stylo 5-6 mm. longo; staminibus circiter 100. 


A shrub or tree with elliptic leaves 2-2.7 times as long as wide, inconspicuously 
veined. The very prominent glands in the inflorescence serve to distinguish this 
species from all others except E. polyadena Berg, which as noted in the key to 
species has differences in pubescence and considerably larger flowers, as well as 
a different geographical range. Buds 5 mm. long; larger calyx-lobes 2-2.5 mm. 
long; flowers up to about 18. 

Ecuador: El Recreo, Prov. Manabi, y S. Lat., H. F. A. Eggers 
15787 (GH; US 1361943, type), Eggers 14957, Apr. 29, 1897 (F). 
Locality uncertain: Fl. H[?uayaquil.j no. 408, [anno] 1803 (Herb. 
Barbey-Boissier) . 

The last specimen cited above is of uncertain origin but may 
have formed a part of the Flora Huayaquilensis, of which some 
specimens were distributed, probably by Pavon. The collector may 
have been Tafalla, but even this is uncertain. Another specimen 
of this same species, from the Moricand herbarium and now at 
Geneva (G), is labelled "Eugenia sp nova, Peru," and in another 
hand, "Pavon." This is probably one of a considerable series which 
Moricand received from Pavon in 1827, and which included plants 
from various parts of America, collected in part by Pavon himself, 
and in part by others. From the state of preservation of this 
particular specimen, from its morphology and from its degree of 
maturity, it appears that the plant originally formed a part of the 
collection cited above, which is attributed tentatively to Tafalla. 
The Flora Huayaquilensis is known to include species from various 
parts of what is now Ecuador, including the higher Andes (e.g., 
Myrteola microphylla var. microphylla} , so that it is impossible to 
make a definite statement that the above specimens of Eugenia 
pustulescens came from any particular part of Ecuador. In view of 
the known modern localities where the species grows, however, it 
seems most probable that the early collection or collections came 
from the coastal lowlands of Ecuador; this, if true, suggests that the 
species may justifiably be excluded from the known flora of Peru. 

Eugenia quadrijuga McVaugh, sp. nov. 

Frutex vel arbor, plusminusve appresse pubescens; foliis ellipticis, 7-11 cm. 
longis obtuse acuminatis; venis utroque latere circiter 10, utrinque elevatis; 
nervo medio supra impresso; venulis tenuiter reticulatis; racemis abbreviatis; 
floribus 2-8 pedicellatis, pedicellis minute hispidulis, pilis brevissimis adscenden- 
tibus obsitis; alabastris 4.5-5 mm. longis; bracteolis longiusculis, 1.5 mm. longis 
persistentibus nee connatis; disco 1.5-2.5 mm. lato; staminibus circiter 60-75. 

A shrub or tree up to 15 meters high; for discussion of the variability and 
relationships of this species, see above under E. discreta; calyx-lobes up to 2.5-3 
mm. long; style 6-8 mm. long; fruit globose or pyriform, probably about 1 cm. 
in diameter. 


Peru, Loreto: Pumayacu, between Balsapuerto and Moyobamba, 
G. King 3153 (A; F; G; GH; US). Brazil, Acre: near mouth of Rio 
Macauhan (tributary of Rio Yaco), Lat. 9 20' S., Long. 69 W., on 
terra firma, Aug. 21, 1933, B. A. Krukoff 5594 (NY; US 1664166, 
type; Y); same locality, Krukoff 5415 (NY; US). Seringal S. 
Francisco, E. Ule 9661 (G; US). 

Eugenia quebradensis McVaugh, sp. nov. 

Arbor vel frutex, tomentosus, vel gemmis ramulisque novellis sericeo-veluti- 
nis, pilis sordidis usque ad 1 mm. longis obtectis; foliis lanceolatis, 5.5-7.5 cm. 
longis, 5-6-plo longioribus quam latioribus; racemis abbreviates, floribus majus- 
culis sessilibus glomeratisque; bracteolis ignotis, ut videtur deciduis; calycis lobis 
rotundatis, 3-3.5 mm. longis; disco 6 mm. lato. 

A tree or shrub, with the aspect of some species of Psidium, well marked by 
the tomentum, the narrow rigidly coriaceous leaves and the large sessile flowers; 
hypanthium 4-5 mm. long and wide, with 4 strong winglike angles; style probably 
10-15 mm. long; stamens about 300. 

Known only from the type, which is from a botanically little- 
studied area. 

Peru, Lambayeque: Prov. Chiclayo, quebrada del Rio Sana, 
monte seco, Dec., 1928, N. Esposto s.n. (USM, type). U. S. Nat. 
Mus. Neg. 4466. 

Eugenia quinqueloba McVaugh, sp. nov. 

Arbor pilosa, ramulis dichasiisque pilis appressis vel adscendentibus usque 
ad 0.8 mm. longis obtectis; pilis pellucidis, luminibus rufidis; foliis subsessilibus, 
orbiculatis vel late ovatis, supra flavido-viridibus vernicosisque; dichasiis 3-7- 
floris; alabastris 5-6 mm. longis; floribus 5-meris; calycis lobis inaequalibus, 
rotundatis, majoribus 3, 3 mm. longis; disco 3-4 mm. lato. 

A tree 4-6 meters high, unique in its sessile and often suborbicular rigid 
leaves up to 7.5 cm. long and wide, its 5-merous flowers in small dichasia, and its 
isolated position in the Department of Lima. Dichasium up to about 3 cm. long; 
style 6 mm. long; stamens 75-100; petals white or yellowish (Ferreyra); fruit 
probably ellipsoid, 1 cm. long or more. 

Peru, Lima: Prov. Huarochiri, arriba de San Bartolome", monte 
bajo, elev. 2,900-3,000 meters, Nov. 5, 1954, R. Ferreyra 10417 
(MICH, type); same locality, Ferreyra 10424 (MICH). 

A very distinct species of uncertain relationships. Were it not for 
the 5-merous flowers, the characters of the embryo, placentation and 
inflorescence would align E. quinqueloba perfectly with the group of 
species that I take to represent the genus Anamomis Griseb. Surely 
E. quinqueloba has little affinity to the genus Myrcianthes, a small 
eugenioid group of eastern warm-temperate South America, in 


which its 5-merous flowers would place it according to Berg's system. 
Possibly it is somewhat more closely related to the monotypic 
Chilean genus Reichea Kausel, which has, however, mostly solitary 
flowers, a very short erect radicle, and apparently somewhat different 
placentation. (In E. quinqueloba the ovules are about 20 in each of 
the two locules, radially and externally directed from a short, 
centrally affixed placenta; the testa of the [somewhat immature] 
seed is free and membranaceous, the cotyledons distinct, fleshy, 
plano-convex, the radicle accumbent and at least half as long as the 
cotyledons; for the corresponding details in Reichea, see Kausel in 
Lilloa 13: 129-130. 1947.) I have assigned the present species to 
Eugenia in spite of the 5-merous flowers, pending revision and ex- 
amination of the generic characters of the whole group of Andean 
species which apparently comprise the major portion of the so-called 
genus Anamomis, as well as the characters of the Chilean genera 
Reichea and Myrceugenella. See also some remarks above, under 
Eugenia ferreyrae. 

Eugenia scalariformis McVaugh, sp. nov. 

Arbor vel frutex; foliis adultis subglabris, 33-35 cm. longis, 4-plo longioribus 
quam latioribus; venis utroque latere 25-30; nervo marginali aperto, valido, vix 
arcuato; petiolo 4 mm. crasso, 12-15 mm. longo; ramis annotinis ut videtur flo- 
riferis, floribus pedicellatis, ?glomeratis; alabastris ut videtur 2 cm. longis; bracteo- 
lis 10 mm. longis, appressis, deciduis; calycis lobis rotundatis, imbricatis con- 
cavisque, interioribus 15 mm. longis; staminibus ut videtur circiter 500, antheris 
1.5-1.8 mm. longis linearibus. 

A shrub or tree with very long elliptic or oblanceolate acuminate leaves with 
numerous scalariform lateral veins; the inflorescence is densely felted with short 
flaccid pale brown mostly dibranchiate hairs; hypanthium 8-angled, 10 mm. long; 
disc 10 mm. wide, glabrous; style 2.5 cm. long; petals 25-32 mm. long. 

Superficially resembles Eugenia tumulescens, from which it may 
be distinguished by the smooth petiole, the more numerous veins, 
the pedicels, which are longer than the hypanthium, and the much 
larger flowers. 

Peru, Loreto: Stromgebiet des Maranon, Santiago Miindung am 
Pongo de Manseriche, G. Tessmann 4328 (G, type). F.M. Neg. 

Eugenia schunkei McVaugh, sp. nov. 

Arbor, subglabra (gemmis rufo-strigosis; calycis lobis bracteolisque minute 
ciliatis; disco sparse piloso); foliis (ramulorum terminalium validorum) 14-23 
cm. longis acuminatis; venis utroque latere circiter 10, apicem versus sensim 
extenuatis, superioribus venam marginalem formantibus; racemis abbreviatis; 
alabastris 12-15 mm. longis pyriformibus; bracteolis 1-2.5 mm. longis persistenti- 


bus nee connatis; calycis lobis 8-10 mm. longis oblongis, apicem versus incras- 
satis cucullatisque, intus glabris, ad florendi tempus reflexis; staminibus 300 vel 
ultra, antheris 1.5 mm. longis linearibus in alabastro erectis. 

A tree 5 meters high with oblong-lanceolate leaves often 3.5 times as long as 
wide and the marginal vein evident in the distal one-third of the blade; racemes 
with 5-6 approximate decussate pairs of flowers, often several together on short 
spurlike excrescences; disc 4 mm. wide; style 12 mm. long. 

A species which is evidently allied to E. feijoi Berg and others 
which are sometimes referred to the genus Catinga, but with very 
much larger flowers than any other known species of this group. 

Peru, Loreto: Rio Mazan, Quebrada Luno, on river bank, elev. 
110 meters, Feb., 1935, Jose M. Schunke 184 (A; US 1459093, type). 
Univ. of Mich. Neg. 450. 

Eugenia stipitata McVaugh, sp. nov. 

Arbor, hispidula; foliorum venis 6-10, arcuatim adscendentibus, inter se 
arcuatis sed venam marginalem vix formantibus; racemi ramis oppositis, 1-floris, 
vel ramis dichotomis 3 (vel raro 7) -floris, flore in dichotomiis stipitato; pedicellis 
1-floris 10-20 mm. longis, longitudinaliter acutangulatis striatisque; bracteolis 
1-2 mm. longis linearibus deciduis; hypanthio turbinato; disco quadrangulato, 
piloso; calycis lobis rotundatis 4, imbricatis, intus appresse pubescentibus, ad 
florendi tempus reflexis; germine 4-loculari, ovulis in quoque loculo circiter 10, 
ut videtur biseriatim angulo loculorum interno affixis. 

A species of uncertain systematic position, apparently without 
any close relatives. The branching of the inflorescence appears to 
be unique among the American species of Myrtaceae. The structure 
of the ovary suggests that of the Subtribe Pimentinae, but the seeds 
(known in subsp. sororia only) are definitely eugenioid in structure 
although relatively more numerous than is usual in the Eugeniinae. 
The species occurs in two well-marked populations, either one of 
which would probably be described as an independent species if it 
were found geographically isolated. These populations, described 
below as subspecies, are readily separated by the characters given 
in the key, but they have so many qualitative characters in common 
that they are surely to be considered as conspecific. 

1. Folia ovata vel late elliptica, 5-6 cm. lata, 1.8-2.3-plo longiora quam latiora, 
venis supra impressis, pagina inferiore pilis erectis acutisque et usque ad 0.5 
mm. longis crebro obsita; pedicelli plerumque in medio fere, vel infra medium 
bibracteolati; stylus 7-8.5 mm. longus, glaber; calycis lobi 4-5 mm. longi. 

subsp. stipitata. 

1. Folia elliptica, 2.5-4.5 cm. lata, 2.2-3.3-plo longiora quam latiora, venis 
supra vix manifestis, nee impressis; pagina inferiore maturitate glabra, vel 
venis solum hispidulis, vel pilis minutissimis 0.1 mm. longis crebro obsita; 
pedicelli 3-5 mm. ultra hypanthium bibracteolati; stylus 5-6.5 mm. longus, 
basi pilosus; calycis lobi 2.5-3 mm. longi subsp. sororia. 


Eugenia stipitata McVaugh, subsp. stipitata. E. stipitata 
McVaugh, as to type. 

More markedly hispidulous than the subsp. sororia, with larger and broader 
leaves, more conspicuous veins and larger flowers. Stamens 100-150. The fruit 
is unknown. 

Peru, Loreto: San Antonio, on Rio Itaya, Killip & Smith 29469 
(NY; US), Williams 3397 (F). La Victoria, Williams 2787 (F). 
Mishuyacu, near Iquitos, forest, elev. 100 meters, Jan., 1930, G. 
King 788 (F 624179, type; NY; US). Brazil, Amazonas: Mun. 
Humayta, near Livramento, B. A. Krukoff 6591 (NY; US). Univ. 
of Mich. Neg. 464. 

Eugenia stipitata McVaugh, subsp. sororia McVaugh, subsp. 

Differs from subsp. stipitata as noted under that taxon and in the key. Sta- 
mens about 75; fruit oblate, velutinous, about 1.5 cm. across; seeds 6-15, reniform, 
3-7 mm. long, the embryo completely undivided or the cotyledons slightly sepa- 
rated at the chalazal end; testa membranaceous. 

Peru, San Martin: Juanjui, Alto Rio Huallaga, elev. 400 meters, 
forest, Oct., 1934, G. King 3834 (F; GH, type; US). Tarapoto, 
Williams 5486 (F), 5667 (F). Bolivia, ?Beni: Junction of rivers 
Beni and Madre de Dios, H. H. Rusby 597 (F; MICH; NY; US). 
Brazil, Amazonas: Near mouth of Rio Embira, B. A. Krukoff 4859 
(US). Colombia, Meta: Villavicencio, Bro. Apollinaire Myrt. no. 2 
(US). Univ. of Mich. Neg. 474. 

Eugenia tenuimarginata McVaugh, sp. nov. 

Arbor vel frutex, foliis ramulisque adultis glabris, racemis minute appresse- 
pubescentibus; foliis 10-14 cm. longis obtuse acuminatis, supra scabrido-papil- 
losis; venis utroque latere 6-8, supra paullum elevatis; racemis abbreviatis; 
bracteolis connatis persistentibus; alabastris 8-9 mm. longis; calycis lobis rotun- 
datis, late imbricatis, intus glabris, marginibus fragilibus, hyalinis, tenuioribus; 
lobis interioribus majoribus 5-8 mm. longis latisque; disco 6 mm. lato; stylo 9-10 
mm. longo. 

A tree or shrub, with elliptic, elliptic-ovate or -obovate leaves about twice 
as long as wide, the veins rather prominent beneath, and the surfaces somewhat 
papillose. The flowers are large, probably always in clusters on old wood; stamens 
about 250; anthers 1-1.2 mm. long. 

Peru, Loreto: Mouth of Rio Santiago, on high land, G. Tess- 
mann 4213, anno "1924" (fragment, F; G, type; NY). Univ. of 
Mich. Neg. 437. 


Eugenia tumulescens McVaugh, sp. nov. 

Frutex tumulescens, ramis crassiusculis, cortice atro-rufido squamuloso 
secedente; racemis tomentulosis; foliis petiolatis, ellipticis oblongisve, 20-38 
cm. longis, 2.5-5-plo longioribus quam latioribus; nervo medio utrinque elevato; 
venis utroque latere 20-30; petiolo rimoso 10-18 mm. longo, ut videtur 3-4 mm. 
crasso (exophloeo suberoso soluto); ramis annotinis floriferis, racemis abbreviatis; 
bracteolis 4-5 mm. longis oblongis, ad florendi tempus erectis; disco 3.5-4 mm. 

A shrub said to form mounds 70 cm. high, with elongate stiff veiny leaves 
which are acuminate at tip and narrowed or rounded at base to the corky-thickened 
petioles. Flowers 6-8, small for the size of the leaves; hypanthium 3 mm. long, 
obtusely 8-ridged and expanded into the gamosepalous calyx-base 1 mm. long; 
style 10-12 mm. long; stamens 250-300, the anthers 2.4-2.6 mm. long; fruit 
long-ovoid, salmon-yellow (according to Froes), probably 3-5 cm. long. 

Brazil, Amazonas: Rio Cauabury, between Rio la and Rio 
Maturaca, E. G. Holt & E. R. Blake 438, Nov. 3-7, 1930 (US); 
Porto Curucuhy, Rio Negro, terreno arenoso alto, beira do rio, 
R. Froes 21106, Oct. 6, 1945 (MICH, type; NY). 

Eugenia valvata McVaugh, sp. nov. 

Frutex vel arbor usque ad 5 m. alta, glabra; foliis plerumque 5-6 cm. longis 
ellipticis obtusis, venis inconspicuis; racemis 1-4 abbreviatis, floribus pedicellatis, 
in glomerulis umbelliformibus conglobatis; bracteolis persistentibus 1-1.5 mm. 
longis ciliatis; calycis lobis ciliatis subdeltoideis, ut videtur valvatis, inter se ad 
florendi tempus longitudinaliter rumpentibus. 

An intricately branched shrub or small tree 3-5 meters high, glabrous except 
the bristly receptacular disc and the ciliate perianth-lobes and bracteoles; leaves 
elliptic, coriaceous, (1.5-) 2-3.5 cm. wide, (3-) 5-6 cm. long, about twice as long 
as wide, obtuse or obscurely acuminate at tip, rounded or gradually narrowed 
at base to the petiole 3-5 mm. long; margins somewhat pale-cartilaginous and 
revolute; midvein sulcate above, prominent beneath; lateral veins 6-10 pairs, 
very slender, scarcely apparent in mature leaves; marginal vein 2-3 mm. from 
margin, about as strong as the laterals and somewhat arched between them; upper 
surface of blade smooth, lustrous, the lower surface paler; glandular dots numerous, 
small, apparent in young leaves but hardly at all on mature foliage; inflorescence 
an abbreviated axillary raceme, or usually a cluster of 2-4 racemes from each axil, 
the axis of the raceme 3-6 mm. long, bearing 5-7 approximate, decussate pairs 
of flowers on pedicels (5-) 11-15 mm. long and up to 0.8 mm. wide at the some- 
what compressed apex; bracts thin, deltoid or ovate, reddish brown, 0.7-1.5 mm. 
long; bracteoles lanceolate or ovate, persistent, appressed to the base of the 
hypanthium, 1-1.5 mm. long; hypanthium subcylindric, 1-1.5 mm. in diameter, 
1.5-2.5 mm. long, somewhat enlarged distally, then abruptly widened into the 
base of the spreading-ascending rounded-deltoid calyx-lobes; lobes membrana- 
ceous, markedly convex without in the bud, ciliate at tips or in the distal half only, 
united by the proximal margins and separating at the time of anthesis by longi- 
tudinal splits up to 1-1.5 mm. long; globe of petals in the opening bud 1.5-2 
times as long as the calyx; disc 3 mm. wide; style 5-6.5 mm. long; stamens about 


75, about as long as the style, the anthers 0.6-0.8 mm. long; petals white or pinkish, 
about 5 mm. long; ovary bilocular, the ovules 12-20 in each locule, attached 
radially to the central septum. 

A species which is distinctive because of the umbelliform clusters 
of flowers and the splitting of the calyx; the calyx-lobes show 
scarcely a trace of any imbricate condition even in the youngest 
bud, and are well separated by irregular short breaks below the 
sinuses by the time the flower opens. Fruit of this species is not 
definitely known, but Hitchcock's collection includes a single de- 
tached fruit, with persistent bracteoles and calyx-lobes, which is 
apparently dark in color, globose and about 1 cm. in diameter. 

Ecuador, Chimborazo: Canon of the Rio Chanchan, about 5 km. 
north of Huigra, elev. 5,000-6,500 feet, moist forested valleys in the 
afternoon fog-belt, May 19-28, 1945, W. H. Camp E-3280 (MICH) ; 
canon of the Rio Chanchan, open deforested slope with small patches 
of scrub in the draws, directly above Huigra, elev. 7,000 feet, May 
29-31, 1945, W. H. Camp E-3512 (MICH, type); Huigra, elev. 
1,200 meters, A. S. Hitchcock 20733 (US). Canar: Santa Rosa de 
Cafiar, J. N. Rose 22655 (US); between Tambo and Suscal, north 
rim of the valley of Rio de Canar, Camp E-2757 (MICH). 

Eugenia variareolata McVaugh, sp. nov. 

Arbor vel frutex, rufo-tomentosus; foliis maximis, 30-45 cm. longis obovatis 
oblanceolatisve, supra lucidis, subtus minute ceroso-papillosis glaucisque; margin- 
ibus cartilagineis; venulis utrinque reticulatis; racemis abbreviatis; bracteolis 
2.5-3 mm. longis persistentibus; calycis lobis intus glabris, ovatis, 7-9 mm. longis, 
ad florendi tempus reflexis, deciduis; disco 4.5 mm. lato. 

A tree or shrub, the branchlets, inflorescence and petioles closely tomentose 
with coarse flexuous lustrous tangled dark red-brown hairs up to 0.5 mm. long, 
a few hairs persistent on the veins of the lower leaf-surface; leaves obovate or 
oblanceolate, 10-15 cm. wide, 30-45 cm. long, about 3 times as long as wide, 
rounded from above the middle to a broad short acumen, and narrowed from near 
or above the middle to near the base, where abruptly contracted, subcordate, with 
low rounded basal lobes; cartilaginous margin (about as thick as the marginal vein 
on the upper surface) passing abruptly into the flat summit of the petiole, which 
is 3-4 mm. thick, 10-15 mm. long; midvein convex and sometimes shallowly 
sulcate above, 1.5-2 mm. wide at base, prominent beneath; lateral veins 12-15 
pairs, ascending, convex but sometimes also impressed above, prominent beneath; 
marginal vein (5-) 10-20 mm. from margin, evident but appearing as a series of 
asymmetric loops formed by the laterals; minor veins prominulous outside the 
marginal vein, including 2 successively smaller submarginal veins forming rather 
symmetrical arches connected to the inner veins by small right-angled veinlets; 
veinlets on both surfaces forming angular areoles of varying sizes; upper surface 
of blade smooth, lustrous, green or drying brown, the lower surface reddish brown 
or paler, glaucous, minutely waxy-papillose; both surfaces obscurely gland-dotted; 


inflorescence an abbreviated axillary raceme (or 2-3 racemes from the same axil), 
the axis up to 5 mm. long, bearing as many as 4 approximate, decussate pairs of 
flowers on pedicels 1-2 mm. thick, 10-12 mm. long; bracts scarious, ovate or 
subrotund, 1-2 mm. long; bracteoles broadly ovate to suborbicular, broad at base, 
persistent, 2.5-3 mm. long; hypanthium bluntly 4-angled, 2-3 mm. long, hemi- 
spheric to subglobose; calyx-lobes membranaceous, thin-margined, ovate, bluntly 
pointed, 4-6 mm. wide near base, 7-9 mm. long, glabrous within, reflexed at 
flowering time and finally dehiscent; disc quadrangular, 4.5 mm. wide, its center 
1.5 mm. wide, deeply depressed (1 mm.), with red-hirsute margin; style glabrous, 
more than 10 mm. long; ovary bilocular, the ovules about 35 in each locule, 
attached radially to the central septum. 

A most distinctive species, of which unfortunately neither buds, 
complete flowers nor fruits are known. 

Colombia, Meta: Villavicencio, elev. 450 meters, Jan., 1856, 
J. J. Triana 14 (BM, type; COL; NY). Univ. of Mich. Neg. 485. 

Eugenia versicolor McVaugh, sp. nov. 

Arbor, ramulis racemisque puberulis, pilis pallide rufis, crispiusculis, brevibus, 
obtectis; foliis 6-13 cm. longis acuminatis, venis utroque latere 6-10; racemis 
abbreviatis; bracteolis persistentibus, connatis; calycis lobis late rotundatis, intus 
appresse pubescentibus, 1.5-2 mm. longis; stylo 7-9 mm. longo; folii pagina 
inferiore ferruginea vel cinerea, pilis crebris minutissimis nitidis obtecta. 

A tree to 15 meters high with elliptic leaves 2-3 times as long as wide, rela- 
tively few veins, and umbelliform clusters of flowers; the lustrous upper surface 
of the leaves contrasts markedly with the rusty or ashy color of the lower surface; 
buds 4.5-6 mm. long; disc 2-2.5 mm. wide; stamens about 200. 

Compared in the key with E. heterochroma Diels, and perhaps 
related to that species or to E. ferreiraeana Berg, the type of which 
came from near the mouth of the Rio Negro, Brazil. 

Brazil, Amazonas: Basin of Rio Solimoes, Mun. Sao Paulo de 
Olivenca, basin of Creek Belem, high forest, terra firma, Oct. 26- 
Dec. 11, 1936, B. A. Krukoff 8910 (MICH; US). Colombia, 
Amazonas : Trapecio amazonico, Loretoyacu River, elev. 100 meters, 
Nov., 1945, R. E. Schultes 6959 (US, type). 


1. Pedicels 4-5 mm. long; flowers 2-4 pairs in short racemes; buds completely 
closed, apiculate, 7-8 mm. long; plants appressed-hispidulous with no long 
silky hairs P. clausa McVaugh. 

1. Flowers sessile or subsessile, in sessile clusters subtended by sterile bracts; 
buds, if closed, 12 mm. long; plants variously hirsutulous or silky-pilose 
in the inflorescence. 

2. Buds 12 mm. long, completely closed, whitened and felted with long ap- 
pressed hairs; stamens about 500; leaves 4-6.5 cm. wide. 

P. inflata McVaugh. 


2. Buds 6-7 mm. long or less, silky-pilose or hirsutulous, the calyx-lobes free 
at tips; stamens 125-150 (number unknown in P. pinnata); leaves various. 

3. Leaves hirsutulous beneath, with hairs about 0.5 mm. long, the veins with 
some longer hairs up to 2 mm. long; hypanthium 2-2.5 mm. across; style 
4.5-6 mm. long P. pinnata L. 

3. Leaves with minute hairs 0.2 mm. long on the lower surface, the young 
branchlets and leaves having also some hairs up to 4 mm. long; hypanthium 
3.5-4.5 mm. wide; style 9-11 mm. long P. duplipilosa McVaugh. 

Plinia clausa McVaugh, sp. nov. 

Arbor vel frutex, ramulis foliis novellis petiolisque ochraceo-hispidulis; 
alabastris albidis, clausis, apiculatis, 7-8 mm. longis; racemis usque ad 3 mm. 
longis, 4-8-floris, floribus pedicellatis; hypanthio supra germen circiter 4 mm. 
producto; staminibus 200-250. 

A tree or shrub with nearly glabrous, elliptic and narrowly acuminate leaves 
7-10 cm. long, 2-2.3 times as long as wide; the flowers are in short racemes with 
conspicuous membranaceous bracts and bracteoles 2.5-4 mm. long; stamens 
arising from a broad zone occupying most of the distal half of the bud, their bases 
intermixed with short silky hairs, the inner surface of the hypanthium glabrous 
below this. 

Peru, Loreto: Soledad (lower Rio Itaya, near Iquitos), July, 1925, 
G. Tessmann 5287 (NY, type). 

Plinia duplipilosa McVaugh, sp. nov. 

Arbor, ramulis foliisque novellis pilosis, pilis albidis rectis tenuibusque, usque 
ad 4 mm. longis, obsitis; glomerulis dense sericeo-pilosis; ramulis petiolisque, et 
folii paginae inferioris venis minute pubescentibus, pilis erectis 0.2 mm. longis 
immixtis; foliis 10-13 cm. longis acuminatis; glomerulis 4-floris, sessilibus, bracte- 
atis; bracteis 4-seriatis, inferioribus sterilibus; bracteolis 5-6 mm. longis, pilosis 
ciliatisque; alabastris 6.5 mm. longis, calyce 4-dentato, dentibus deltoideis 1.5 
mm. longis; hypanthio supra germen 2.5-3 mm. producto; stylo 9-11 mm. longo; 
staminibus 125-150. 

A tree with elliptic-ovate leaves about 2.5 times as long as wide and 6-10 
pairs of lateral veins; the rather large silky flowers are in clusters of 4 in leafless 
axils on old wood, the clusters subtended by 4-ranked sterile bracts; buds con- 
cealed by straight hairs 2.5 mm. long; calyx and hypanthium glabrous within, 
the limb at maturity nearly quadrangular, 5-6 mm. on a side, the receptacular 
cup 3.5-4.5 mm. wide. 

The flowers and inflorescence in Klug's collection are very like 
those of the type, but the leaves are longer and narrower, more 
prominently veined beneath, and with up to 15 pairs of veins. 
Apparently Cuatrecasas' no. 7092, from the lowlands of eastern 
Colombia, is also conspecific; the flowers in this collection have more 
elongate bracts and free calyx-tips, but are otherwise much like 
Peruvian specimens. 


Peru, Loreto: Yurimaguas, elev. 135 meters, dense forest, Aug.- 
Sept., 1929, Killip & Smith 28007 (NY; US 1461669, type); Mishu- 
yacu, near Iquitos, elev. 100 meters, forest, G. Klug 1155 (US). 

Plinia inflata McVaugh, sp. nov. 

Arbor hispidula, pilosa etiam, pilis eburneis 1-1.5 mm. longis, tenuissimis, 
obsita; glomerulis ut videtur 4-floris, pilis longis concretis dealbatis; foliis 9-15 
cm. longis, 4-6.5 cm. latis acuminatis; bracteis sterilibus 3-4-jugis, 4-seriatis; 
alabastris clausis, apiculatis, 12 mm. longis, calyce demum in lobos 4 subaequales, 
10-12 mm. longi, longitudinaliter direpto; hypanthio supra germen 2 mm. pro- 
ducto; staminibus circiter 500; petalis minusculis 3 mm. longis. 

A tree to 9 meters high, with elliptic-oblong leaves about 2.5 times as long 
as wide, and about 15 pairs of very slender lateral veins; the plant appears to 
differ from others in the same genus in its larger flowers, which are markedly 
whitened by the matted hairs; the splits between the calyx-lobes extend some- 
what deeper than the inner (proximal) margin of the broad hairy staminal ring, 
which is 6-7 mm. wide and extends distally to a line 1.5-2 mm. from the apex 
of the bud; style not seen. 

Brazil, Amazonas: Basin of Rio Madeira, Mun. Humayta, near 
Tres Casas, low terra firma, rare, Sept. 28, 1934, B. A. Krukoff 6365 
(NY, type; US); same locality, Oct. 8, 1934, Krukoff 6525 (NY). 


1. Calyx-lobes 7-9 mm. long and distinct, or the lobes prolonged into narrow 
appendages 6-14 mm. long. 

2. Calyx-lobes 4, ovate, 7-9 mm. long, tomentose without, appressed to the 
bud and covering it, apparently valvate P. ulei Diels. 

2. Calyx-lobes 5, separating irregularly as segments 4.5-5 mm. long and wide, 
each tipped by an erect or spreading narrow foliaceous nearly glabrous 
appendage 1.5-3 mm. wide, 6-14 mm. long P. caudatum McVaugh. 

1. Calyx-lobes 3 mm. long or less, short, broad and rounded, or the buds com- 
pletely closed before anthesis and dehiscing irregularly. 

3. Leaves crenate, obtuse and cuspidate, narrow (1-2 cm. wide, 2-7 cm. long); 
plants finely pubescent; flowers solitary; buds glabrous, about 12 mm. long. 

P. maribense DC. 

3. Leaves entire, or sometimes irregularly undulate, rarely less than 2.5 cm. 
wide and if so acute or acuminate; pubescence and flowers various. 

4. Plants completely glabrous, even to the young vegetative buds; branchlets 
compressed, not angled; leaves 3-5 cm. wide, 7-14 cm. long, narrowed from 
below or near the middle to the acute and mucronate tip; calyx open, slightly 
flaring in bud, the broadly rounded lobes 2 mm. high. . .P. densicomum DC. 

4. Plants with evident and usually abundant pubescence (sometimes on young 
growing parts only); branchlets often angled; leaves and calyx various. 

5. Lateral veins 12-20 pairs, usually impressed above, prominent beneath and 
well differentiated from the smaller intermediate veins; young growth 
heavily pubescent; branchlets 4-angled or sometimes terete; buds closed, 
10-16 mm. long, pointed, not cuspidate; peduncles 1 (very rarely 3) -flowered. 


6. Leaves lanceolate or narrowly ovate, 1.5-3 cm. wide, 3-4 times as long as 
wide, gradually narrowed to the acute or acuminate tip; fruit with 8-15 
irregular longitudinal ridges P. rutidocarpum G. Don. 

6. Leaves elliptic or oblong, 3-6 cm. wide, 2-3 times as long as wide, more 
abruptly narrowed to the obtusely pointed or rounded tip; fruit globose or 
pyriform, smooth P. guajava L. 

5. Lateral veins 6-15 pairs (mostly 10 pairs or fewer), if more numerous then 
slender and scarcely differentiated from the intermediate veins, or the plants 
very sparsely pubescent only; branchlets and flowers various. 

7. Lower leaf-surface hirsutulous, the numerous hairs erect or nearly so, 0.5-1 
mm. long. 

8. Leaves mostly 6-10 cm. long; pubescence of branchlets reddish; buds 10-12 
mm. long, almost completely closed; peduncle 1.5-3 cm. long, 3-flowered. 

P. guineense Sw. 

8. Leaves mostly 5-6.5 cm. long; pubescence tawny yellow; buds 3.5-4 mm. 
long, with 5 broad low calyx-lobes 1 mm. long; peduncle 1 cm. long or less, 
3- to 7-flowered P. fulvum McVaugh. 

7. Lower leaf-surface glabrous or sparingly pubescent or strigose. 

9. Leaves small for the genus, mostly 5.5 cm. long or less, often acute at both 
ends; buds 5-7 mm. long, open, the 5 broad low calyx-lobes much shorter 
than the corolla; branchlets terete or sometimes longitudinally channeled. 

10. Dichasia 3-flowered, with very slender divaricate branches; calyx-lobes 1 mm. 
long or less; marginal vein of leaf strongly arcuate, 1.5-4 mm. from the 
margin at the points where the arches join the lateral veins; glands not 
apparent even in young leaves P. pedicellatum McVaugh. 

10. Flowers solitary; calyx-lobes more than 2.5 mm. long; marginal vein about 
1 mm. from margin, not strongly arcuate; leaves gland-dotted on both sides 
at least when young P. arayan (HBK.) Burret. 

9. Leaves larger, 5-10 cm. long or even longer; buds 10-16 mm. long, closed 
at apex; branchlets various. 

11. Branchlets quadrangular, wing-angled; buds shortly apiculate; lateral veins 
10 pairs or fewer; leaves usually with very numerous dark raised glands 
beneath; Amazonian lowlands, widely distributed P. acutangulum DC. 

11. Branchlets compressed, sometimes with low rounded ridges, not wing-angled; 
buds with linear or subulate apiculum 2-5 mm. long; lateral veins 10-15 
pairs; leaves with numerous small open glandular depressions of varying 
sizes, on both surfaces; Pacific slopes, Tumbez P. rostratum McVaugh. 

Psidium caudatum McVaugh, nom. nov. Psidiopsis moritzi- 
ana Berg, Linnaea 27: 351. 1856, non Psidium moritzianum Berg, 
I.e. 359. Calycolpus moritzianus (Berg) Burret, Repert. Sp. Nov. 50: 
57. 1941. 

As far as I am aware, no one has been able to point out any 
distinguishing feature of Psidiopsis except the prolonged tips of the 
calyx-lobes. This is certainly noteworthy but scarcely indicative 
of a profound evolutionary hiatus between the one species which is 
so marked, and the other species of Psidium. 

Psidium fulvum McVaugh, sp. nov. 

Arbor vel frutex, fulvo-hirsutulus; foliis integris ovatis, 5-6.5 cm. longis 
obtusis; venis utroque latere 8-10; pedunculis usque ad 1 cm. longis, 3-7-floris; 


alabastris 3.5-4 mm. longis; calycis lobis 5, 1 mm. longis, 2-2.5 mm. latis; disco 
5-angulato, 3-4 mm. lato; germine 3-loculari, sporophoris axillaribus peltatis 

A shrub or tree, rather conspicuously and densely tawny-hirsute with hairs 
up to 0.5 mm. long; leaves broadly ovate, 1.5-2 times as long as wide; flowers 
small; style glabrous, not seen fully expanded, probably peltate; stamens about 

Peru, Amazonas: Chachapoyas, A. Mathews ["derniere collect."] 
(BM, type; G). Univ. of Mich. Neg. 484. 

Psidium pedicellatum McVaugh, sp. nov. 

Arbor, puberula; foliis integris immaturis 3-5.5 cm. longis, brevi-acuminatis; 
venis utroque latere 7-10; pedunculis usque ad 1.5 cm. longis, ramis filiformibus 
5-6 mm. longis divaricatis; dichasiis 3-floris; alabastris 5 mm. longis; calycis 
margine arcuato-undulato 5-lobato, lobis 0.8-1 mm. longis; disco glabro, 2.5 mm. 
lato; germine 2-loculari, sporophoris axillaribus peltatis 20-ovulatis. 

A tree 12 meters high with scant puberulence, the leaves broadly elliptic-ovate 
and 1.5 (-2) times as long as wide; glands not apparent even in young leaves; 
dichasia usually 3-flowered, the flowers occasionally solitary; flowers small, the 
style 4-5 mm. long, the stigma subpeltate; stamens about 150. 

Ecuador, Santiago-Zamora : Along Quebrada Honda, vicinity of 
Rancho Achupallas, elev. 2,500-2,700 meters, along river, Oct. 10, 
1943, J. A. Steyermark 54571 (F 1391169, type). Univ. of Mich. 
Neg. 472. 

Psidium rostratum McVaugh, sp. nov. 

Arbor vel frutex, pubescens vel strigosus; foliis integris 10-14 cm. longis 
acutis vel apice obtusis; venis utroque latere 10-15; pedunculis 1-floris, axillaribus, 
vel infimis ex nodis efoliatis, vel rhachi abortiva, oriundis; alabastris 12-16 mm. 
longis clausis, longe apiculatis; calyce intus dense strigoso; disco 12-14 mm. lato. 

A shrub or tree to 10 meters high, rather sparingly pubescent, with irregularly 
ovate to oblong or even obovate leaves about twice as long as wide, and irregularly 
glandular-pitted on both surfaces; flowers large, the style 10-13 mm. long, the 
stigma peltate; stamens probably about 300. 

Peru, Tumbez: Mountains east of Hacienda Chicama, in decidu- 
ous bushwood, elev. 900-1,000 meters, Feb. 19-24, 1927, A. Weber- 
bauer 7648 (F 571783, in flower; F 571784, type, in bud). Univ. of 
Mich. Negs. 471 (type), 494. 


1. Plants glabrous and subherbaceous; stamens 12 or fewer; northern Peru to 
Venezuela M. oxycoccoides (Benth.) Berg. 

1. At least the young branches densely pubescent, or if exceptionally the whole 
plant glabrous, an erect shrub with 30 or more stamens. 


2. Leaf-margins strongly revolute; leaves densely strigose or setose beneath; 
stamens more than 30; northern Peru and Ecuador. [M. acerosa (Berg) 
Burret; M. microphylla (Humb. & Bonpl.) Berg, var. microphylla.] 

2. Leaf-margins not or scarcely revolute; leaves glabrous beneath or sparingly 

3. Stamens 20 or fewer; leaves mostly broadest below the middle, with a tend- 
ency to become bullate at least in age; plants with extensive rhizomes, the 
flowering branches prostrate or erect, often less than 20 cm. high (up to 
1 meter); bracteoles 1.5-3 mm. long; calyx-lobes 1.5-2 mm. long. 

4. Leaves narrowly lanceolate, 1.5-2 mm. wide, 6-7 mm. long, bullate from the 
first, the midvein not apparent beneath. 

M. vaccinioides var. carabaya McVaugh. 

4. Leaves ovate, 2.5-5 mm. wide, 4-8 mm. long, bullate in age or not at all, the 
midvein apparent at least in young leaves. 

M. vaccinioides (HBK.) Berg, var. vaccinioides. 

3. Stamens 30-65; leaves broadest near the middle or but slightly below it, not 
bullate or only exceptionally so, the midvein apparent beneath; erect shrubs 
mostly 1-2 meters high; bracteoles 3.5-6 mm. long; calyx-lobes 2.5-3.5 
mm. long. [M. microphylla var. glabrata Berg; M. weberbaueri Diels.] 

Myrteola vaccinioides (HBK.) Berg, var. carabaya McVaugh, 
var. nov. 

Frutex, strigoso-hispidulus, rhizomatus; ramis floriferis 6-15 cm. longis erectis; 
foliis anguste lanceolatis, subtus bullatis, nervo medio haud manifesto; bracteolis 
2-2.5 mm. longis; staminibus circiter 15. 

Further revisionary study in this genus, or examination of ad- 
ditional material of the present plant, may well indicate that this 
is an independent species. Unquestionably, however, it is akin to 
M. vaccinioides, which it much resembles in habit, in stamen-number, 
and in all vegetative characters except the leaf-shape. 

Peru, Puno: Prov. Carabaya, June-July, 1847, H. A. Weddell 
4667 (P, type). Bolivia, La Paz: Unduavi, en bosques, elev. 3,300 
meters, Feb. 12, 1907, 0. Buchtien 647 (NY). Univ. of Mich. 
Neg. 428. 

Publication 806